• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

5x5 Rescramble Best Strategy

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,682
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
I have done 3x3 Rescramble in 2015 and 4x4 Rescramble in 2018.
3x3 Rescramble (with inspection, but participants mutually agreed not to use inspection) is already an event in Red Bull Cup since the last 3 years. Also 3x3 Match the scramble (without inspection) has been ongoing in the Speedsolving weekly comps for over a decade.

Currently, I am practicing a bit of 5x5 Rescramble (multiple looks, not one look) just for fun and noticed a few things that I would like to share. There are two approaches that I feel are contesting, and I am in a dilemma on which approach is the best.

Method 1:
1)To do centers in normal way (start from solved cube and match all 6 centres), and match the 6 centres (regard less of how much messed up outer layer gets).
The drill is the same as the 5x5 speedsolve drill, first one centre, then the opposite centre, then 2 adjacent centres on the E layers, and then last 2 centres.
2) Solve the tredges+corners using comms,
and then do inverse comms for tredges and corners, to match the scramble.

pros: no need of doing centre comms, which take time to execute as there are lot of slice moves. 6 centers can be done in 6 look with practice.
cons: match last 2 centres is tough, we have to use some comms to fix the 6th centre, since the 6th centre does not automatically match, like what happens in a 5x5 speed solve.
wings and midges+corners needs to be solved and then unsolved(unsolved to match the scramble), which cost, ~22 extra comms.


Method 2:
2) Complete BLD method via inverse comms of all 5 piece types (will be quite slow to do ig)

pros: number of looks required will be constant, about 35-43 (one look per comm)
Doing 1 look possible too, since it is BLD method, but one look will cost memo time, and few DNF attempts.

cons: center execution will take a lot of time
there might be lot of inverse comm mistakes
doing the parity first and then the solve is tricky, we have to trace 2 piece types for that (corners and wings). The trace to see parity wastes about 10-20 seconds.


P.S. : The current 5x5 Rescramble one-look record is held by yuanzi with a time of 11:11.30. I do not think multiple looks 5x5 Rescramble has been tried out by anyone before.

I will try reconstructing Rescramble solves using both these strategies to see which one takes less moves. I am not an expert in making simulations of cube states, and I think doing simulation for this experiment is not that trivial.

Also, if you find any other efficient way of matching the scramble, do let me know in this thread.
 
Last edited:

Mike Hughey

Administrator
Staff member
Joined
Jun 7, 2007
Messages
11,303
Location
Indianapolis
WCA
2007HUGH01
SS Competition Results
YouTube
Visit Channel
Just a minor correction (perhaps it was a typo on your part) - 3x3x3 match the scramble in the weekly competition is and has always been WITH inspection. That rule was set up originally by Arnaud van Galen when he first introduced the event to the competition. Also I am surprised by your description of the Red Bull rules; I thought that inspection was specifically not allowed by their rules.

The 5x5x5 rescramble does seem like quite a challenge. It seems like if I were to do this, my first inclination would be to first do centers normally as in your method 1 (but relying more on commutators for the last 2 centers), then solving everything else with commutators (like your method 2).
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,682
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
Just a minor correction (perhaps it was a typo on your part) - 3x3x3 match the scramble in the weekly competition is and has always been WITH inspection. That rule was set up originally by Arnaud van Galen when he first introduced the event to the competition. Also I am surprised by your description of the Red Bull rules; I thought that inspection was specifically not allowed by their rules.

The 5x5x5 rescramble does seem like quite a challenge. It seems like if I were to do this, my first inclination would be to first do centers normally as in your method 1 (but relying more on commutators for the last 2 centers), then solving everything else with commutators (like your method 2).
Oops, looks like I have been doing 3x3 Match the scramble wrong for the last 4 years. No wonder I was able to match with top scores. I use BLD method for 3x3 match the scramble.
Red bull keeps changing the rules about inspection, for the 2021 online qualifiers they have kept inspection, but they are not fixed on the format. For example, last year there was inspection from the Rubik's connected app display, which is different from normal inspection.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
omg, why is info about rescramble so hard to find?! I have been trying 5x5 rescramble using CFOP for weeks and I always run into a parity (??) case where I either need to swap just two corners or rotate just one face 90°, both of which are impossible. I finally figured out a convoluted work-around but there must be a better way!

To be more specific, my method is to try and reverse engineer a 5x5 reduction method (with free slice). The first four centers are simple, but the last two require a supercube approach, but even then there is sometimes a parity issue which requires switching two of the center corner pieces using conjugates (with a 4x4 supercube safe alg). Then I pair all the edges (easy enough unless there is edge parity--but there is a 5x5 supercube safe edge parity alg), form the cross, and use F2L to match all but the top layer. Then I form the "yellow cross" and try to match the corresponding edges. But, more than 50% of the time, this can't be done, so I try to put the corner pieces in the correct position and it turns out that two of them need to be swapped. I try things like t- and j-perm, but these rotate the top center 90 degrees, which leaves me with a completely matched cube except for a 90 degree rotation of the top center!! So I found a supercube safe alg that swaps two corner pieces (and rearranges the center in an adjustable way with conjugates) and messes up four edges. Then I fix the edges, reform the cross, resolve F2L, and the "corner parity" isn't there anymore and I can match the scramble (using supercube safe PLL algs). Seems like there ought to be a better way!
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,682
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
omg, why is info about rescramble so hard to find?! I have been trying 5x5 rescramble using CFOP for weeks and I always run into a parity (??) case where I either need to swap just two corners or rotate just one face 90°, both of which are impossible. I finally figured out a convoluted work-around but there must be a better way!

To be more specific, my method is to try and reverse engineer a 5x5 reduction method (with free slice). The first four centers are simple, but the last two require a supercube approach, but even then there is sometimes a parity issue which requires switching two of the center corner pieces using conjugates (with a 4x4 supercube safe alg). Then I pair all the edges (easy enough unless there is edge parity--but there is a 5x5 supercube safe edge parity alg), form the cross, and use F2L to match all but the top layer. Then I form the "yellow cross" and try to match the corresponding edges. But, more than 50% of the time, this can't be done, so I try to put the corner pieces in the correct position and it turns out that two of them need to be swapped. I try things like t- and j-perm, but these rotate the top center 90 degrees, which leaves me with a completely matched cube except for a 90 degree rotation of the top center!! So I found a supercube safe alg that swaps two corner pieces (and rearranges the center in an adjustable way with conjugates) and messes up four edges. Then I fix the edges, reform the cross, resolve F2L, and the "corner parity" isn't there anymore and I can match the scramble (using supercube safe PLL algs). Seems like there ought to be a better way!
Interesting, I have not understood your workaround fully yet.

Can you explain it through an example attempt on video maybe?
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
By now I’ve perfected my (very inefficient and slow, but 100% effective) method. It takes me ~30 minutes to match a 5x5 scramble if there’s no corner parity. (Corner parity adds 10 minutes or so!) The tricky parts are the last two centers and the corner parity. I hope to post a video soon(ish) detailing my workaround.
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,682
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
By now I’ve perfected my (very inefficient and slow, but 100% effective) method. It takes me ~30 minutes to match a 5x5 scramble if there’s no corner parity. (Corner parity adds 10 minutes or so!) The tricky parts are the last two centers and the corner parity. I hope to post a video soon(ish) detailing my workaround.
I am confused. Why does corner parity take additional 10 minutes? Should it not be just an algorithm?


Also last 2 centers should not be tricky right, since we just use commutators to get the pieces in place.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
I'm no speed solver (I'm lucky to solve a 5x5 in under 4 minutes), but my corner parity algorithm messes up four edges and F2L, so I'd estimate it takes me 10 minutes to fix everything and finish the rescramble (I've never timed myself, though).

You are right about commutators for the last two centers, but there is a parity case that (unless I'm mistaken) doesn't show up in solving a 5x5 supercube. For example, what would you do if you needed to swap the top right and bottom right center pieces while leaving the other centers unchanged?
 
Joined
Dec 19, 2016
Messages
743
Location
Home
WCA
2015FOXC01
YouTube
Visit Channel
You are right about commutators for the last two centers, but there is a parity case that (unless I'm mistaken) doesn't show up in solving a 5x5 supercube. For example, what would you do if you needed to swap the top right and bottom right center pieces while leaving the other centers unchanged?
I've not tried 5x5 rescramble, but couldn't you just do another commutator for that parity case? I'd assume you could just get another center of the same color taken from another side.


I'm no speed solver (I'm lucky to solve a 5x5 in under 4 minutes), but my corner parity algorithm messes up four edges and F2L, so I'd estimate it takes me 10 minutes to fix everything and finish the rescramble (I've never timed myself, though).
As for the corner parity alg, would it be possible to just use like a T perm or something to solve the corners, then use standard parity algs to solve the edges? Seems like a deceptively simple solution if it works, but something feels off about it, I'll have to give it a shot for myself later.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
Yes, I suppose you’re right about the last two centers—the alg I have for that case probably does exactly what you suggest!

The T perm (or any other corner switching PLL alg) will rotate the center 90°, though!

I do suggest trying a 5x5 rescramble. I think it’s a lot harder in practice than it sounds in theory.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
Update: I’ve got the 5x5 rescramble down to 18-20 minutes—with a ridiculous amount of improvement possible—if there’s no corner parity. Still trying to find efficient ways to resolve this parity case….
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,682
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
Update: I’ve got the 5x5 rescramble down to 18-20 minutes—with a ridiculous amount of improvement possible—if there’s no corner parity. Still trying to find efficient ways to resolve this parity case….
Can you try and get one of your attempts on camera, so that a lot of people interested can figure out your techniques and your time splits to do 5x5 rescramble successfully.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
Yeah, sorry—I know I said, like, 3 months ago that I’d post a video, but I just haven’t gotten around to it. I still plan to do it soon! I’ll tell you right now that my splits will not be impressive, but that just means incredible improvement on my ~18 minute (sans parity) time is possible.
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
More to come, but my splits (with no corner parity) lately have been about 5 min for centers, 7 min for edges, and 3 min for the rest. So 15 min rough average, but I’d wager that that could easily be cut in one-fifth by you speed demons!
 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
I finally have a video. Sorry for the poor production value! It ended up being an easier one with no corner parity. If you want to know, the scramble was F Lw' Fw Bw' Dw' L Dw' L2 Uw2 Lw2 D U L' R2 Uw B' Lw' D R' D' Lw F U' D' Bw F' R2 Bw Dw' B' Rw' B2 D' B Dw' F U2 Lw D Fw Uw' Dw Bw R2 Bw2 Fw Dw B L F' L2 D2 R' Bw L2 Rw B' Dw2 Fw Lw'.

I also attached my supercube safe algorithms to solve the last four edges, and my algorithm to swap UR center corner with LR center corner (remember this also swaps top center edge with right center edge, so a set-up move may be needed first). Also, the alg I mentioned to solve corners in the next-to-last center is U Rw U' Lw' U Rw' U' Lw

For reference, my steps are: opposite centers, two adjacent centers, edges of last two centers, corners of last two centers, edges (last edge parity needs to be explained in another video), cross, first "two" layers, orient top edges, solve top corners (dreaded corner parity discussed in a future video), solve top edges. Sounds like a lot, but I'm confident some of you can get this down to 3 minutes if there's no corner parity.

algs.jpeg
 
Last edited:

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
Here's a video explaining my solutions to the three parity cases (last center parity, last edge parity, and corner parity). Any improvements on these methods (especially for corner parity) would be greatly appreciated! Algs are below.

Last center parity: Fw' U' Rw Fw Rw' U' Rw Fw' Rw' U2 Fw U'
Corner parity: U Rw2 Bw2 Rw F Rw' Bw2 Rw F' L2 r U' b2 U L' F' L U' b2 U L' F M x

 

mattmc2

Member
Joined
Feb 24, 2022
Messages
12
Location
Columbus, OH
Here's a video explaining my solutions to the three parity cases (last center parity, last edge parity, and corner parity). Any improvements on these methods (especially for corner parity) would be greatly appreciated! Algs are below.

Last center parity: Fw' U' Rw Fw Rw' U' Rw Fw' Rw' U2 Fw U'
Corner parity: U Rw2 Bw2 Rw F Rw' Bw2 Rw F' L2 r U' b2 U L' F' L U' b2 U L' F M x

Edit: When talking about the last edge parity case, I should’ve said the standard alg rotates the center 180° and swaps the top and bottom center edge pieces (which is why you’d like them to be the same color).
 
Top