# 7x7 centers by reduction

#### rubiksfriend

##### Member
Actually, I solve a 1x1x5 band, insert two outside 1x1x5 bands, then the remaining inner 1x1x5 centers.

Last edited:

#### Stefan

##### Member
I tried this a few more times and the general version is still slow for me (about 33% more time than my "normal" way). But for the last two centers, I often do this, and like it.

#### Joël

##### Member
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 exactly what I have been doing intuitively. Works for me.

#### Joël

##### Member
I tried your idea once Stefan... I have to get used to it a lot, but it's not a bad idea... I was already used to making one face at a time.

#### mrCage

##### Member
Hmmm ....

I really need my (taxfree this month) salary, and order those cubes. I see thay are constantly re-stocking, not much. Just little by little.

My last 2 checks were as follows

5x5x5:
127, 136

5x5x5(black):
104, 139

6x6x6:
103, 215

7x7x7:
81, 206

Cheers!!

- Per

#### AvGalen

I only did a couple of solves (weekly competition) and centers were harder then expected (I suck at commutators during speedsolving. It takes me > 1 minute to "develop" one). I basically tried about 10 things for centers afterwards and the one I liked best was "spiraling out" by doing edge, corner/edge-pair, corner/edge-pair, row:

For 7x7x7 this means you do this (think of the centers as coordinates A1 to E5)
A1 B1 C1 D1 E1
A2 B2 C2 D2 E2
A3 B3 C3 D3 E3
A4 B4 C4 D4 E4
A5 B5 C5 D5 E5

1 dot in the middel (C3) = start
-----
-----
--x--
-----
-----

Attach 1 central edge (B3) = edge
-----
-----
-xx--
-----
-----

Attach 1 central edge + 1 central corner (B2+C2) = corner/edge-pair
-----
-xx--
-xx--
-----
-----

Attach 1 central edge + 1 central corner (D2+D3) = corner/edge-pair
-----
-xxx-
-xxx-
-----
-----

Attach 1 central edge + 2 central corners (D4+C4+B4) = row
-----
-xxx-
-xxx-
-xxx-
-----

Attach 1 outer edge + 2 outer wings (A4+A3+A2) = edge
-----
xxxx-
xxxx-
xxxx-
-----

Attach 1 outer edge + 2 outer wings + 1 outer corner (A1+B1+C1+D1) = corner/edge-pair
xxxx-
xxxx-
xxxx-
xxxx-
-----

Attach 1 outer edge + 2 outer wings + 1 outer corner (E1+E2+E3+E4) = corner/edge-pair
xxxxx
xxxxx
xxxxx
xxxxx
-----

Attach 1 outer edge + 2 outer wings + 2 outer corners (E5+D5+C5+B5+A5) = row
xxxxx
xxxxx
xxxxx
xxxxx
xxxxx

This spiraling allows for a lot of freedom during the solve and you get a lot of "just attach" situations. It also works for even cubes, you just have to add a start piece because those are nog always there.

And most importantly, if you see something nice; Use it! Many times you can get about 50% of the centers done in about 10 moves. Even if you end up breaking/repairing some of that later you save a lot of moves and get better look-ahead.

I also tried and liked "bars/squares inside out"

Start
-----
-----
--x--
-----
-----

Attach 1 edge
-----
-----
--x--
--x--
-----

Attach 1 more edge
-----
--x--
--x--
--x--
-----

Attach 1 row
-----
-xx--
-xx--
-xx--
-----

Atttach 1 more row
-----
-xxx-
-xxx-
-xxx-
-----

Attach 1 edge
-----
-xxx-
-xxx-
-xxx-
-xxx-

Attach 1 more edge
-xxx-
-xxx-
-xxx-
-xxx-
-xxx-

Attach 1 row
xxxx-
xxxx-
xxxx-
xxxx-
xxxx-

Attach 1 more row
xxxxx
xxxxx
xxxxx
xxxxx
xxxxx

And this also works for even sized cubes

#### mrCage

##### Member
Hi

I have read in many posts already about making those 1x1x5 bars. How is that done (in some detail). What i figure is that they are built on some other layer and then inserted. Does this still work with only 2 unsolved faces to solve ? etc ....

- Per

#### AvGalen

Hi

I have read in many posts already about making those 1x1x5 bars. How is that done (in some detail). What i figure is that they are built on some other layer and then inserted. Does this still work with only 2 unsolved faces to solve ? etc ....

- Per
Yep, you build the bar/line/row on another face and then attach it to the correct face. This doesn't work really well for the last 2 faces. It is basically done like in this video, only with 5 pieces instead of 3. There are no algs, only intuition

#### alexc

##### Member
Hi

I have read in many posts already about making those 1x1x5 bars. How is that done (in some detail). What i figure is that they are built on some other layer and then inserted. Does this still work with only 2 unsolved faces to solve ? etc ....

- Per
Yep, you build the bar/line/row on another face and then attach it to the correct face. This doesn't work really well for the last 2 faces. It is basically done like in this video, only with 5 pieces instead of 3. There are no algs, only intuition
It works for me for the last two faces! I first build the middle 1x5 which is pretty easy because you have room to work with. Then I add on the inner 1x5's which I find is also easy. Then I put in the outer two. The last step is the most difficult. Here's how I do it. I try to connect the outer + center with the two corner centers, at least. (I will connect an oblique too if I can.) The point is I don't need to finish the 1x5 before inserting it. So, say you have the two corners and the outer + center connected. Insert this onto the correct face. Then, to solve the remaining obliques use Rw U' l' U Rw' U' l and its reflection. Repeat for the final 1x5.

#### mrCage

##### Member
Hi

I'm getting the feeling that this bar-building business is really only a camouflaged commutator. Nothing wrong in that of course

- Per

#### AvGalen

Off course it still works, just not as well as for the first 4 centers. For the first 4 centers it is just "position, move out of the way, restore". For the last 2 centers it is basically a commutator or another type of alg

For the last two centers I have been using this order (only done about 10 solves so probably not the best way to do it)

1: Inner 3x3 centers (B2-D2,B3-D3,B4-D4. This is easy)
2: Most other centers (very easy) using just a couple of moves
3a: Solve remaining outer corners (A1,E1,A5,C5 with wide-sune(s) (Rw U Rw' U Rw U2 Rw')
3b: Solve remaining outer center-edges (C1,A3,E3,C5) using an "alg"/intuition like Rw S U2 S' U2 Rw'
3c: Solve remaining oblique(s) using a simular "alg" (or just intuition as alexc

#### AvGalen

[checker off]and i forgod an klosing brakket two. Its plane too si dad mine english sugs. im affraid i wil ceep makeing simular misstakes indefiantly.[chekker on][chegger on][checkker on][checker on] => auto-corrected to [/checker off]

What about the actual content of my posts? Do you like any of these reductions?

#### Stefan

##### Member
Yeah, sorry, you just write that quite often, and since it's actually a correct word with a quite different meaning...

Apparently like you, I don't like building five 1x5 blocks. I'm doing it somewhat like you, first I build the inner 3x3 centers square like I'm used to from the 5x5 cube, then I add two opposite 1x3 center tredges, then the remaining two 1x5 blocks. I'll try your spiraling out as well as your above 1/2/3abc, which apparently the fast guys use, too (those I've watched).

Btw, I think the main reason I started this thread was people asking for help because they had trouble solving the last two centers at all. So it was an idea to solve it with a technique they already knew (from solving edges).

#### AvGalen

[spelling]I don't mind people correcting my English, but after writing such a detailed post I thought it was weird that your reaction was: "simular"

Somehow I keep messing up the same words over and over
plane/plain
indefinitely (I am not the only one)
then/than
to/too
of/off/offcourse
simular/similar
[/spelling]
I'm doing it somewhat like you, first I build the inner 3x3 centers square like I'm used to from the 5x5 cube, then I add two opposite 1x3 center tredges, then the remaining two 1x5 blocks.
I don't think this is "like" me. It is exactly how I described "bars/squares inside out" (unless you do the inner 3x3 differently).

And I didn't understand your topic was about the last two centers. Could you give a scramble and a solution?

#### mrbiggs

##### Member
Off course it still works, just not as well as for the first 4 centers. For the first 4 centers it is just "position, move out of the way, restore". For the last 2 centers it is basically a commutator or another type of alg
I've been using the exact same method you posted earlier for the first four centers, but I used it on the last two as well. I solve the inner 3x3x3 like I would on a 5x5x5. To triple the oblique centers I've been doing something like (I don't have the cube with me so there might be a typo):

3R' U2 F 2R U2 2R' F' 3R

where 3R is the third R slice, 2R is the second, etc.

and now there's a matched up triple on the U face, near L which I can insert the same way I do the inner 3x3x3. It's the same pair, move out of the way concept, just done on two instead of three faces. It's not the fastest method in the world, but I only have to do it about twice per solve.

#### Stefan

##### Member
I'm doing it somewhat like you, first I build the inner 3x3 centers square like I'm used to from the 5x5 cube, then I add two opposite 1x3 center tredges, then the remaining two 1x5 blocks.
I don't think this is "like" me. It is exactly how I described "bars/squares inside out" (unless you do the inner 3x3 differently).
Yes, I do the inner 3x3 differently. For example if these are U and F:

Code:
???
O?O
???

??O
?OO
??O
Then M' U M builds a 2x3 centers block on F. And other stuff, too.

And I didn't understand your topic was about the last two centers. Could you give a scramble and a solution?
It kinda wasn't, at least I didn't present it that way. Rather than typing this, I'd show it on video, but that might have to wait until after US Open. But it's not that special/hard anyway, is it?

Last night I tried the spiraling on the 6x6 and realized I had been violating your "if you see something nice; Use it!" principle. For example after extending the inner 2x2 to a 2x3, I always extended to a 2x4, even if the additional 2x1 was actually a 3x1 so I could've extended to 3x3 rather than 2x4. Now I think that was quite stupid, as it wastes an opportunity and makes further extensions harder.

#### UMichSpeedCubist

##### Member
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:

Code:
ABBBC
DEEEF
DEEEF
DEEEF
GHHHI
Step 1 builds the E square, step 2 builds the B/D/H/F triples.
I've been doing this for over 1 year on gabbasoft's 7x7. I had the idea pretty early. And it doesn't show much promise. It's something to add to the arsenal, but the recognition is pretty tough even after I've practiced it for a long time. There are cases when the a center triple is already done, I can place using the "outer-5x5" route and that saves a lot of time.

But in general, it's fairly risky for me and I've switched back to doing full columns now. Making the inner/central 3x3 block is very fast for me since I know almost every optimal case, but then the outer +centers might take a lot of turns for me.

Although I have learned some new tricks this week, perhaps it still has merit. Either way, it's probably the most fun way of solving centers on 7x7! The way I do it is 1 entire face at a time, not all the inner/central ones first and then go back and do the outer stuff.... I think I'd lose a bit of focus if I did it your way. Although I might have tried that once or twice. Need to experiment more for sure.

Btw, I do very little 6x6 solving. It might be really nice way of doing those centers as someone else here said earlier.

-Doug

#### Stefan

##### Member
Yes, I like spiraling. Although, I might "change direction" at any point. So a better name might be "growing rectangle staying as close to square as possible". Or maybe that's too long.