7x7 centers by reduction

[Stefan]

An idea to solve the 7x7 centers: Reduce to 5x5 centers.

Step 1) Build 3x3 centers on all sides like you would on a 5x5.
Step 2) Build "center tredges" like you'd build tredges on a 5x5.
Step 3) Solve centers like 5x5 centers.

Hope that's clear, and hope this hasn't been mentioned before. Well, I think it's good and I don't hear Per cursing, so that might be a good sign.

Edit to make it clearer: Steps 1+2 make the centers of each side look like this:

ABBBC
DEEEF
DEEEF
DEEEF
GHHHI

Step 1 builds the E square, step 2 builds the B/D/H/F triples.

 

[masterofthebass]

This is almost impossible. Have you tried this? The only way to get the outer "tredges" is to use commutators, which isn't very fast.

 

[Mike Hughey]

Have you tried it? How fast were you able to do it?

I'm just awful at 7x7x7 centers (it looks like others have the same problem), so anything like this that might help would be welcome. I'm going to go home and try it tonight for sure.

My current approximate splits (for the 2 solves I've tried so far since solving it BLD):
First 4 centers: 11 minutes
Last 2 centers: 4 minutes
Edge matching: 6 minutes
3x3x3 final solve: 1 minute

 

[Gprano]

That's exactly how i did it when i tried the 777 on gabbasoft.
I matched the tredges like the edges, breaking the 555 centers, exchanging with an unsolved tredge, restoring the 555 centers.
That worked, but i found it very slow...

 

[Derrick Eide17]

i came up with something like this for the FINAL 2 centers. i think it works well and its how i am going to solve them i think (for now) so like create 3 1x5 bars on 1 face and the other 3 1x5 bars is pretty much almost done. then finish filling in list few centers around outside with commuators (3 at most)
then solve last corner centers just like 5x5 centers.

 

[masterofthebass]

Erik gave me the idea of somewhat doing this. Instead of building 1x5 blocks, you just build the inner 3x3 block. Then you extend that block with a couple of 1x3s and then fill in the 2 outer 1x5s I've been using it and to a degree of success

 

[Jack]

Erik gave me the idea of somewhat doing this. Instead of building 1x5 blocks, you just build the inner 3x3 block. Then you extend that block with a couple of 1x3s and then fill in the 2 outer 1x5s I've been using it and to a degree of success

That was actually the method I came up with when first trying 6x6 and 7x7 computer cubes. It seems like it could be about as fast as the 1x5 block method.

 

[Jason Baum]

Erik gave me the idea of somewhat doing this. Instead of building 1x5 blocks, you just build the inner 3x3 block. Then you extend that block with a couple of 1x3s and then fill in the 2 outer 1x5s I've been using it and to a degree of success

That's basically what I've been doing on 6x6x6: build the inner 2x2 block, add a 1x2 block, add two 1x3 blocks, and add the final 1x4 block. It's pretty nice. I like it better than just buildilng four 1x4 blocks.

 

[Derrick Eide17]

Erik gave me the idea of somewhat doing this. Instead of building 1x5 blocks, you just build the inner 3x3 block. Then you extend that block with a couple of 1x3s and then fill in the 2 outer 1x5s I've been using it and to a degree of success

lol i told erik one time that thief! nah im prety sure he came up with something like that first.

*ponders*

 

[TimMc]

My slow method:
- 3x3 centers on each face
- Extend to 4x4 centers on each face (keep them aligned so that you've a slot to work with)
- Complete Down and Upper layers by any means
- Solve the remaining 4 * 4x1 center edges
- Solve the remaining 4 * 5x1 center edges

*shrug*

Tim.

 

[masterofthebass]

On 6x6 I just build the 1x4 blocks, since they are pretty quick. On good solves I get in the 1:30s right now. For the 7x7, my centers are up in the 3:30s I need to work on speed during centers, as they can be much faster.

 

[Erik]

Erik gave me the idea of somewhat doing this. Instead of building 1x5 blocks, you just build the inner 3x3 block. Then you extend that block with a couple of 1x3s and then fill in the 2 outer 1x5s I've been using it and to a degree of success

lol i told erik one time that thief! nah im prety sure he came up with something like that first.

*ponders*

Hey hey, I just gave him the idea, I didn't say it was MY idea, cause in fact it clearly isn't.

 

[Stefan]

This is almost impossible. Have you tried this? The only way to get the outer "tredges" is to use commutators, which isn't very fast.

No, you can do it just like edges with the usual five moves for pair-replace-otherPair. And you even have more freedom/choices for the replace part.

I have tried it, but not often enough to compare it to building 1x5 blocks (which I've used more often). And my cubes are unlubed and stiff, so I'm pretty slow anyway.

I'm also thinking about just using it for the last two centers. I start those by solving their inner 3x3 centers anyway. Which btw has the benefit that these centers would need pure inner slice commutators while the outer centers can be solved partly with double layer commutators.

 

[RobinBloehm]

This is almost impossible. Have you tried this? The only way to get the outer "tredges" is to use commutators, which isn't very fast.

No, you can do it just like edges with the usual five moves for pair-replace-otherPair. And you even have more freedom/choices for the replace part.

I have tried it, but not often enough to compare it to building 1x5 blocks (which I've used more often). And my cubes are unlubed and stiff, so I'm pretty slow anyway.

I'm also thinking about just using it for the last two centers. I start those by solving their inner 3x3 centers anyway. Which btw has the benefit that these centers would need pure inner slice commutators while the outer centers can be solved partly with double layer commutators.

Yes, that's what I do for the last center, too. Swapping two inner center is really annoying, so I build the 3x3 block in the middle, usually being able to add a few pieces while doing that to have almost 3 1x5 blocks instead of only the 3x3, then sometimes swapping two full bars of the remaining 1x5 blocks, and now solving up to 4 single pieces with double layer commutators or maybe two pieces at once if they are on the same slice.

 

[aznblur]

For last 2 centers, its not hard to create 1x5 strips in the middle. But the outer 1x5 strips are harder, and normally there is one piece that is wrong in each of the strips. So for me, I only have to swap 2 center pieces.

But this is an interesting idea, I shall try it out.

 

[mrCage]

OK Stefan, I will curse a little bit

In all honesty i would instead use as much as possible the same easy technique as for the 5x5x5 centers. For the harder cases use easy commutators. The center commutators are as easy to understand as cotton (!!!) after minimum practice with them :-D Who said a hybrid method is not allowed ??

- Per

 

[masterofthebass]

Guys, for the last 2 centers, I have a maximum of 2 commutators to do, which end up being the oblique centers Everything else can be solved relatively easily. I start off building the center 1x5 block on once face. I then build a 1x5 in the next outermost layer, and then the next, giving me half of the face solved. Then, I pair up the inner 1x3 block just like 5x5 and then attach the 2 outer obliques to it. Then, I attach the outer xcenters with the outer T and then se commutators for the last 2 obliques. It goes really smoothly and has very few commutators.

 

[ExoCorsair]

Meh, just do it freestyle with commutators for the last two.

 

[Mike Hughey]

I just wanted to mention I did try Stefan's original idea as I said I would, and it certainly worked fine, so I really don't understand why Dan said it was impossible. It's a completely successful functioning method. No need for commutators at all - if you use the AVG method of pairing tredges and use the same method here for centers, the same kinds of moves work just fine (except that sometimes you have to do slice moves to get the pieces in place). (Maybe this method just doesn't work with the bigcubes.com edgematching method? It really works fine with AVG.)

However, that being said, my most recent solve prior to trying it was 13 minutes, but my first attempt using Stefan's original suggestion took 23 minutes. (centers: 8 minutes vs. 18 minutes) So it was outrageously S-L-O-W. But it worked just fine.

I would guess that with some practice this could become a much more reasonable method. But I doubt it could ever become really super fast. It's just faster to line up 1x5 blocks, I think.

I will probably keep trying this off and on as I hopefully get more decent at 7x7x7; I'd like to think that someday I can get this method sub-10. It certainly should be possible to do that. The neat thing about that is that then most of the cube is solved with the AVG edgematching method.

 

[masterofthebass]

I thought it was impossible because I couldn't figure out any way to pair up the "edges" Personally, i think it's a very slow idea, so I really didn't put too much effort into figuring out a way to do it.