Times: What is my biggest problem?

[Carson]

I average somewhere around 1:15 for solves. My best times are around :55 and my worst times (excluding completely screwing up and having to redo steps) are around 1:35.

I currently use Fridrich, but haven't learned all the algorithms... currently this is my method:

* Cross

* For Corner/Edge insertion, I learned all of the algorithms for cases in which the initial cross color on the edge piece to be inserted is facing to the left when placed in the upper layer above where it is to be placed. I use these algorithms and their mirrors. So basically, I know algorithms for all cases except when the edge piece has the cross color facing up.

* My LL is pretty weird. I think it is a 4 Look LL... I orient edges to form top cross; use 1 of 7 OLL algorithms for cases in which edges are already oriented; Apply the 'A' perm once or twice based on situation to permute corners; and finally apply 1 of 4 perms that affect only corners.

For my LL: I apply from 1-3 algos for orienting edges, 1 for orienting corners, 1-2 for permuting corners, and 1 for permuting edges... for a total of 4-7 algorithms. I realize this is an insane amount of moves for the LL, and I did have a much quicker LL using the beginner method, but hopefully learning more algorithms will help me out.

______________________________________________________________
I just did seven solves... for each solve, I stop the timer after solving the cross, then reset and solved the f2l before stopping it again, and finally restarting it and solving the LL. These times aren't completely accurate since I am starting and stopping, but they should give everyone an idea of what percentage of my solves are spent on what sections of the cube.

1
Cross:6.25
CE Insertion:36.71
LL:38.23

2
Cross:13.59
CE Insertion:28.92
LL:26.15

3
Cross:6.51
CE Insertion:31.72
LL:24.99

4
Cross:13.61
CE Insertion:27.66
LL:24.09

5
Cross:14.81
CE Insertion:25.99
LL:20.49

6
Cross:11.76
CE Insertion:23.79
LL:26.19

7
Cross:15.19
CE Insertion:36.48
LL:24.93
__________________________________________________________
This averages to
Cross:11.67
CE Insertion:30.18
LL:26.44
__________________________________________________________
My own thoughts:
My cross is horrid, I have spent a lot of time working on this, but it doesn't seem to get much better. I am hoping I have one of those breakthrough moments one day and everything just becomes clear.

My corner/edge insertion takes forever. A lot of this has to do with recognition, but a lot of it is also that I have to do setup moves for most of the cases so that the pieces are in a configuration for which I know an algorithm. I hope to have the rest of the f2l algorithms down in about a week. Then I should have a better idea of how much of an issue the recognition truly is.

I don't even know what to say about my last layer. I just have a lot of memorization to do I suppose.




What does everyone else think? What is my next step?

 

[PCwizCube]

For your cross, what side do you form it on? The up or down side? I think for you level your times for your cross are decent. Just keep it consistently under 12 second on average.

You learned algorithms for your C/E? Except the cases where the cross color is for the corner is facing up? Yeah, that's a lot of algorithms, which can be a pain in the neck for recognition, memorizing, and execution. You should learn it intuitively first - that will be much easier to learn and improve on. Learn from this tutorial:
Part 1: http://www.youtube.com/watch?v=k-xbcAMfWwM
Part 2: http://ca.youtube.com/watch?v=4GxLM_dZqg4

You are using 4LLL (4 Look Last Layer). Just practice it and your last layer will get better. You should been aiming for 5 seconds or less for each step (4*5 = 20 seconds) NOTE: 4LLL can be done in 8 seconds or less
If you need help on understand 4LLL or the algorithms go to Badmephisto's very helpful tutorials:
2 Look OLL: http://www.youtube.com/watch?v=DTYvklyOpVM
2 Look PLL: http://www.youtube.com/watch?v=S61q3FYVFis

Your cross is okay, learn F2L (C/E) intuitively and practice it a lot, and just practice your 4LLL.

You should be aiming for these times (you could probably get reach this goal in 2-4 weeks):
Cross = 10 seconds
F2L = 25 seconds
4LLL = 20 seconds
Total = 55 seconds

I would recommend getting these times before moving onto your next step, 3LLL (full PLL - 21 algorithms)
Cross = 5 seconds
F2L = 20 seconds
4LLL = 15 seconds
Total = 40 seconds

* My LL is pretty weird. I think it is a 4 Look LL... I orient edges to form top cross; use 1 of 7 OLL algorithms for cases in which edges are already oriented; Apply the 'A' perm once or twice based on situation to permute corners; and finally apply 1 of 4 perms that affect only corners.

You mean edges right?

 

[Carson]

For your cross, what side do you form it on? The up or down side? I think for you level your times for your cross are decent. Just keep it consistently under 12 second on average.

I solve cross on the up side, and flip it to the down side as part of my final couple moves to finish cross. Doing this, I am able to spot a corner edge pair immediately about 50% of the time. I have tried solving cross on down, but almost always get confused and end up sticking an edge in the wrong orientation. That is my goal however.

* My LL is pretty weird. I think it is a 4 Look LL... I orient edges to form top cross; use 1 of 7 OLL algorithms for cases in which edges are already oriented; Apply the 'A' perm once or twice based on situation to permute corners; and finally apply 1 of 4 perms that affect only corners.

You mean edges right?

Yup, edges... Is this a typical 4 Look LL? I hear 2, 3, 4 Look LL used a lot, but was never able to find exactly what defines them.

 

[PCwizCube]

For your cross, what side do you form it on? The up or down side? I think for you level your times for your cross are decent. Just keep it consistently under 12 second on average.

I solve cross on the up side, and flip it to the down side as part of my final couple moves to finish cross. Doing this, I am able to spot a corner edge pair immediately about 50% of the time. I have tried solving cross on down, but almost always get confused and end up sticking an edge in the wrong orientation. That is my goal however.

I wouldn't worry about too much about your cross being on the up side now because you need to improve your other steps. It's a great idea to solve it on a down side, and to get used to it, just practice it a lot.

* My LL is pretty weird. I think it is a 4 Look LL... I orient edges to form top cross; use 1 of 7 OLL algorithms for cases in which edges are already oriented; Apply the 'A' perm once or twice based on situation to permute corners; and finally apply 1 of 4 perms that affect only corners.

You mean edges right?

Yup, edges... Is this a typical 4 Look LL? I hear 2, 3, 4 Look LL used a lot, but was never able to find exactly what defines them.

Yes, that is called Fridrich 4LLL. That means solving the last layer in 4 steps. (2 Look OLL + 2 Look PLL) Fridrich 3LLL is solving the last layer in 3 steps (2 Look OLL + Full PLL). Fridrich 2LLL is solving the last layer in 2 steps (Full OLL + Full PLL).

You should stick with 4LLL until you average around 15 seconds for the last layer.

 

[blgentry]

Rubik's Cube Fan has given you some good advice. Especially concerning learning F2L intuitively. I personally would not be learning the F2L algorithms (the 42 cases) at all, and rather focus on the intuitive version, and learn specialized algs for the tough cases later. WAY later.

About the 4 Look Last Layer:

You're not doing a true 4LLL, but you are getting there. A true 4LLL means that in the absolute worst case, you apply 4 algs, and look at the configuration of the cube 4 times. To have a true 4LLL you need the following:

1. 3 algs for orienting edges. This covers all three cases (line, L, and dot) and will guarantee orientation of all edges in one step (by applying the appropriate alg).
2. All 7 OLLs for the cases of all edges solved. You already have this. Good.
3. 7 PLLs so you can guarantee a 2 look PLL.
3.A. Both A perms for permuting corners
3.B. The 4 edges only algs: Both Us, H, and Z.
3.C. One of the algs that does a diagonal corner swap. Macky recommends the Y perm. It's what I used at that stage, and it works well.

That probably only leaves 2 - 4 new algs for you to learn for your 4LLL and then you're set for a while.

Just my opinions and I'm not very fast compared to most here.

Brian.

 

[hawkmp4]

Learning both A perms isn't necessary for a 4LLL. In regards to corners only, each appears to swap two adjacent corners.
By the way, why does everyone want people to learn A perms for 4LLL? its kind of tricky to figure out where everything is going for a new cuber. T perm is much easier- FR and BR corners get swapped. That's it.

 

[Carson]

1. 3 algs for orienting edges. This covers all three cases (line, L, and dot) and will guarantee orientation of all edges in one step (by applying the appropriate alg).

I am currently using:
F R U R' U' F' followed by F (R U R' U') (R U R' U') F' for dot
F (R U R' U') (R U R' U') (R U R' U') (R U R' U') F' for line
and F (R U R' U') (R U R' U') F' for L

I know these are slow, but I didn't think it would matter too much since I intended to learn all the OLL's and this step would be skipped... but now I see that I will be holding off on OLL's for a while, so I guess I should find some better algorithms for this. Any suggestions?

3. 7 PLLs so you can guarantee a 2 look PLL.
3.A. Both A perms for permuting corners
3.B. The 4 edges only algs: Both Us, H, and Z.
3.C. One of the algs that does a diagonal corner swap. Macky recommends the Y perm. It's what I used at that stage, and it works well.
Brian.

I learned the Y perm... I get it now, that saves me from having to do two A perms on occasion. I hadn't considered that. I already know both the A's, U's, Z, and H perms.

Does that mean that I should be able to get down to ~40 seconds without learning anything new?

 

[hawkmp4]

I am currently using:
F R U R' U' F' followed by F (R U R' U') (R U R' U') F' for dot
F (R U R' U') (R U R' U') (R U R' U') (R U R' U') F' for line
and F (R U R' U') (R U R' U') F' for L

I know these are slow, but I didn't think it would matter too much since I intended to learn all the OLL's and this step would be skipped... but now I see that I will be holding off on OLL's for a while, so I guess I should find some better algorithms for this. Any suggestions?

Hm, those are slow, and nowhere near optimal.
F (R U R' U') F' will suffice for the line.
and f (R U R' U') f' will work for the L.
Then, for no edges oriented, you can do
F (R U R' U') F' f (R U R' U') f'
...which is much faster.

 

[blgentry]

I am currently using:
F R U R' U' F' followed by F (R U R' U') (R U R' U') F' for dot

I think you have a typo above, as that doesn't work. Either way, you can use any of the OLLs that have ZERO edges oriented to solve the "dot" case. I like case #17 from:

http://cubefreak.net/OLL.html

(R U R' U) (R' F R F') U2 (R' F R F')

F (R U R' U') (R U R' U') (R U R' U') (R U R' U') F' for line
and F (R U R' U') (R U R' U') F' for L

What hawk said will work MUCH faster than those two algs. Boy are they long!

3. 7 PLLs so you can guarantee a 2 look PLL.
3.A. Both A perms for permuting corners
3.B. The 4 edges only algs: Both Us, H, and Z.
3.C. One of the algs that does a diagonal corner swap. Macky recommends the Y perm. It's what I used at that stage, and it works well.
Brian.

I learned the Y perm... I get it now, that saves me from having to do two A perms on occasion. I hadn't considered that. I already know both the A's, U's, Z, and H perms.

Bingo!

Does that mean that I should be able to get down to ~40 seconds without learning anything new?

A lot of people here seem to have gotten to 40 and lower with a 4LLL. I personally couldn't, but I pretty much started learning the rest of the PLLs as soon as I had the 4LLL "down", so I didn't really give it a chance.

Brian.

 

[mrbiggs]

Does that mean that I should be able to get down to ~40 seconds without learning anything new?

Absolutely. I was under 30 with a 4LLL as I remember, and I could certainly match that at this point (I average low twenties with 3LLL). Nearly all the algorithms for 4LLL are super fast, so if you practice them you're not going to lose more than a few seconds over the 2LLL.

However, you might want to start learning the PLLs immediately after you got the 4LLL down, not because it will have a significant impact on your times now, but because it's nice to not have to suddenly learn 15 new algorithms once you get to a point where you're trying to shave just one or two seconds off your time. But you should realize that learning the new algorithms doesn't serve to decrease your time, it just is going to help you in the future. Getting an 8 move, preplanned cross, a smooth F2L, and good transitions between steps is what lowers your time.

 

[Carson]

Thanks for the input everyone... hopefully I'll break that 1:00 barrier very soon.

 

[Carson]

I average somewhere around 1:15 for solves. My best times are around :55 and my worst times (excluding completely screwing up and having to redo steps) are around 1:35.

Sorry for bringing up an old topic (and for quoting myself... I didn't even realize that could be done really) but I just wanted to do a little follow up. It has been 19 days since I made the above referenced post. I just did a 5 solve average and got the following:

1. 47.41
2. 01:25.88
3. 56.33
4. 55.03
5. 47.47

Average = 52.94
Solve number 2 involved a U perm gone horribly wrong.

I just wanted to say thank you to everyone that provided tips, and also to point out that all I learned to cut down my average by almost 20 seconds, was the Yperm and new algorithms for OLL-Edges. Beyond that, it was just gaining a better understanding of how the cube works and how to better apply what I already knew. So for anyone in a similiar situation... don't think that the only way to get faster is to memorize more stuff, it's really about how to use what you already know.

 

[ShadenSmith]

It's not the size of your boat, but the motion of the ocean.


Of course that saying is generally applied in a different situation.

 

[Neutrals01]

Hmm..for ppl like me, I prefer learning more new algs(because I found a method on how to remember new algs easily, like 10~20 mins per alg)..somehow I counted my speed with an average of 1.7 turns per sec for each solve(include recognition time) for my previous method. My previous method was a LBL and 4LLL, with an average 100 turns (half turn metric). My new method is intuitive f2l(do not need any algs<save the time learning the 42 cases), and 3LLL(2 look orient + PLL) with an average 75 turns (half turn metric). In the end I also end up with a 1.7 turns per sec on average. That is why my time from 57~60s average dropped to 43~46. But I not really use why many ppl say that a 4LLL and LBL method compared to a intuitive f2l + 3LLL method only seconds away, it's like 25 turns diff on average.(maybe because they can do 3~5 turns per sec). I think learning a full PLL will help reduce ur time, reduce ur turns by 10 on average and just need 1 look insted of 2 for permutation(that includes with practice to recognize the pattern as fast as possible,full PLL average 14 turns,2 look Permute average 24 turns).Orientation I think is possible to stick with 2 Look, average 16 turns, compared to full OLL average 10 turns. But I suggest learning the 8 algs of OLL for cases with one dot(edges all incorrectly orient) which is an average 12 turns instead of 22 turns. Hope this helps, although I also know that I am noob in cubing.

 

[coopersacatfilms]

Well, Your last layer can use some work my advice is learn all of PLL and F2L. A good website for F2L is (cubestation.co.uk). Also plan your cross in your inspection not only pieces that are easy to place and worry about the other ones after you place those pieces. Another thing about the cross try to place opposites then do a D turn to place them all. One more thing about the cross, you should do it on the bottom, it takes a little getting used to but it acts as a smoother transition to F2L. Now F2L, Try your best to look ahead, and learn more algorithms. F2L can make or break your solve. Now LL your 2 look OLL is good but PLL and F2L is more important in my opinion. But most of all Practice,practice,practice.

 

[hdskull]

For Orienting last layer, all you need to know before the 7 is.

F (R U R' U') F' for the line.
F (R U R' U')x2 F' or y2 f (R U R' U') f' for the L.
and F (R U R' U') F' f (R U R' U') f' for no edges oriented.

For PLL, if you know As, Us, H, Z, and Y(to solve N perms). You can solve PLL with 2 algorithms or less.

 

[AvGalen]

For Orienting last layer, all you need to know before the 7 is.

F (R U R' U') F' for the line.
F (R U R' U')x2 F' or y2 f (R U R' U') f' for the L.
and F (R U R' U') F' f (R U R' U') f' for no edges oriented.

For PLL, if you know As, Us, H, Z, and Y(to solve N perms). You can solve PLL with 2 algorithms or less.

Actually, you only need 1 of the A's and Y for the corners.
You do need both U's, H and Z for the edges

 

[Mike Hughey]

Actually, you only need 1 of the A's and Y for the corners.
You do need both U's, H and Z for the edges

I prefer using T and Y for the corners. I find that lining up the two corners that match and performing T is more braindead than figuring out the direction for an A.

I just became interested in this because I was learning Jason Baum's square-1 method, and I wasn't able to learn all the PLLs in time for this week's competition in Cincinnati. So I did a 2-look PLL. T or Y for corners is much faster for me than A or Y, due to the quicker recognition time. (I never did a 2-look PLL on the 3x3x3; when I learned Fridrich, I learned all 21 algorithms at once so I didn't have to figure out a 2-look solution. And in the meantime, if I ran into a PLL I didn't know, I just solved the cube the old way I was used to instead.)

 

[AvGalen]

You don't have to figure out the direction of the A-perm. There are only 3 situations for corners:
In Ortega-speak: 4 bars. Don't do anything
In Ortega-speak: 0 bars. Apply Y-perm
In Ortega-speak: 1 bar. Apply clockwise A-Perm.
For more info, look at this video: http://www.youtube.com/watch?v=dnYm2xtVkEA

 

[Mike Hughey]

I guess that does work, yes. And I guess it only makes sense that it would work, too. Now I need to decide whether I prefer A-perm or T-perm on square-1. I think I've gotten so good at T-perm that I might want to stick with it, even though A-perm is a couple of moves shorter. I should probably stick with T-perm for this weekend, though, since it's probably a little too close to the competition to be changing algorithms.