I finally finished my Corner Comms spreadsheet

[DeeDubb]

Check it out here

It's color coded based on the type of comm it is.

I've spent a few months working on this. I still need to memorize about half of it. I'm using Anki to memo both the comm and letter pair.

I created the algorithms mostly myself. I did reference Noah's list to compare my algorithms and change if he had a better one. I prioritized algorithm creation like this: Pure Comms -> Rotationless 9 move -> Rotationless 10 move -> Rotation 9 move -> Rotation 10 move -> columns -> columns with a setup move -> 2 setup move (only two on my whole sheet)

I may revise some of the algs, but if you have any suggestions for better options, let me know.

 

[h2f]

Will you add speffz? I use my own letter scheme and Im not familiar with any other...

 

[mark49152]

It would be great if algdb.net added comms so that we can look them up in one place and compare options.

 

[Hssandwich]

I'm going to (try) to learn these, thanks for making/compiling them!

 

[DeeDubb]

Will you add speffz? I use my own letter scheme and Im not familiar with any other...

I added the targets across the top and on the right side (I can't add on the left because it will throw my letters off).

I'm going to (try) to learn these, thanks for making/compiling them!

Good luck! I recommend trying to learn them on your own and then comparing them.

I started with the FRD target, then RDF, then DFR. That corner seems to be the easiest to start to figure out how comms work.

Also, I did some BLD solves at home yesterday trying to only use corner comms. It seemed like I was able to figure out the correct comm about 90% of the time. So I'm much closer than I thought!!!

 

[h2f]

@DeeDubb Thank you. Most of them I use but sometimes in other rotation.

 

[DeeDubb]

@DeeDubb Thank you. Most of them I use but sometimes in other rotation.

Yeah, actually, sometimes I do the other x rotation or y rotation too. Just depends on how I'm feeling, lol.

 

[h2f]

Have you tried z/z'? I like it.

 

[Hssandwich]

There are a few algs that I changed on a copy of your spreadsheet that I prefer. EG: for CI I prefer [x R': [R' D2 R, U2]].

EDIT: Unless I am misunderstanding the layout, BV should be [R2, D' L2 D] rather than [D' L2 D, R2].

 

[h2f]

I like the L2 D' L2 D L2 and L2 U L2 U' L2 tricks but only second I will implement to my algs.

 

[DeeDubb]

Have you tried z/z'? I like it.

Yeah, I use quite a few z/z' algs, though generally for column cases or various A Perms.

There are a few algs that I changed on a copy of your spreadsheet that I prefer. EG: for CI I prefer [x R': [R' D2 R, U2]].

Yeah, I get that. I originally had that one, but I changed it to avoid the rotation. It's a tough call between a 9 move with a rotation and a 10 move without one. Generally I lean toward the 10 move though. In that case, you can turn the x R' into an l', so I suppose it's about equal. I'm really on the fence about CI/IC.

EDIT: Unless I am misunderstanding the layout, BV should be [R2, D' L2 D] rather than [D' L2 D, R2].

Yeah, that was a mistake. Sometimes I copy and paste wrong. Thanks! Let me know if you find more.

I like the L2 D' L2 D L2 and L2 U L2 U' L2 tricks but only second I will implement to my algs.

I noticed that Noah also avoids doing the D column ones (sometimes he'll z2 to do it with U's). Once i figured out the pattern, it was pretty easy to figure out, so I don't mind the D/D' algs.

 

[unsolved]

I just solved all 88,179,840 corner positions with respect to fixed centers, times all possible (24) rotated states, along with their optimal solutions. I know speed cubers prefer "ease of alg" to fewest moves, but when my computer frees up (currently solving every possible center Alg through depth 12 on the 5x5x5) I'll test your Algs against what I have if you're interested.

 

[bobthegiraffemonkey]

I just solved all 88,179,840 corner positions with respect to fixed centers, times all possible (24) rotated states, along with their optimal solutions. I know speed cubers prefer "ease of alg" to fewest moves, but when my computer frees up (currently solving every possible center Alg through depth 12 on the 5x5x5) I'll test your Algs against what I have if you're interested.

This list is about corner 3-cycles, and move-optimal solutions are already known for all cases. Look up the BH method if you want to see the history of move-optimal 3-cycles. And yes, move-optimal solutions are not necessarily the fastest.

 

[h2f]

I think insted doing y/y' it's better to do U/U'. This is what I want to implement to my list.

 

[DeeDubb]

I just solved all 88,179,840 corner positions with respect to fixed centers, times all possible (24) rotated states, along with their optimal solutions. I know speed cubers prefer "ease of alg" to fewest moves, but when my computer frees up (currently solving every possible center Alg through depth 12 on the 5x5x5) I'll test your Algs against what I have if you're interested.

I'd be interested to see some move optimal algs for certain cases.

I think insted doing y/y' it's better to do U/U'. This is what I want to implement to my list.

That's probably true. I just fear getting confused.

 

[unsolved]

I may revise some of the algs, but if you have any suggestions for better options, let me know.

One of the things I do in my spreadsheets it to create a URL with a formula that will draw the cube on crider.co.uk

It converts the alg in the spreadsheet to the format his site uses to draw cubes. That way, I can see exactly what the alg does.

Example:

http://cube.crider.co.uk/visualcube.php?size=300&pzl=5&alg=U2rR%27fF%27r%27REf%27FU2fF%27E%27rR%27f%27Fr%27R&fmt=png

I see that website is down again, so I can't display the image it would otherwise generate.

Maybe you can do something similar with your spreadsheet, and show the image associated with your algs.

In other news, I am 10% of the way through generating every 11-move "cage alg" that will solve the centers on the 5x5x5. So far, that makes 1.2 billion algs! Here is one where 26 of the 48 free-moving centers are in motion:

2R 3U' 2F' 2R' 3U2 2L 3U 2F 3U 2L' 3U

The idea is to solve the cube to a cage, then blindly apply all of the algs, and execute the alg that solves the greatest number of centers. Do this a few times, and the cage gets solved very quickly (like 8-9 seconds!)

 

[Meneghetti]

Nice job, thanks! It's always good to see other people's list (with the same buffer and lettering scheme)!

One thing I noticed is that sometimes you rotate one way and do LUD moves, instead of rotating the other way to be able to do RUD moves. For example:
SB your way: [y': [L D L', U']]
alternative: [y: [U', R D R']]

Spoiler

Edit: I think that is what h2f was talking about

And I'm not a big fan of the L2 D' L2 D L2 and L2 U L2 U' L2 tricks. I'd rather set these columns up to A9s. For example:
XC your way: [L2 D' L2 D L2, U2]
alternative: [R' U': [U' R U, L2]]

 

[DeeDubb]

Nice job, thanks! It's always good to see other people's list (with the same buffer and lettering scheme)!

One thing I noticed is that sometimes you rotate one way and do LUD moves, instead of rotating the other way to be able to do RUD moves. For example:
SB your way: [y': [L D L', U']]
alternative: [y: [U', R D R']]

Spoiler

Edit: I think that is what h2f was talking about

And I'm not a big fan of the L2 D' L2 D L2 and L2 U L2 U' L2 tricks. I'd rather set these columns up to A9s. For example:
XC your way: [L2 D' L2 D L2, U2]
alternative: [R' U': [U' R U, L2]]

Thanks! Yeah, actually, I think my choice to do y or y' is pretty arbitrary as my ability to make algs improved as I progressed through. For SB, I actually do it with y, but I just haven't changed the alg yet. I'll go back through and fix them later.

As for the column cases, I plan to go back through and try to eliminate the columns that require a setup move then 12 move (as that's 13/14 move) and try to find decent 12-move alternatives. I don't mind the basic ones though.

Great feedback!

 

[DeeDubb]

I'm thinking of making a commutators web page to host all of the algs and make them searchable, plus allow custom letter schemes, and possibly even generate an export file for Anki. I'm pretty sure I have a rough idea how to do it. I'll probably start next week when I have some time.

 

[Miro]

I'm thinking of making a commutators web page to host all of the algs and make them searchable, plus allow custom letter schemes, and possibly even generate an export file for Anki. I'm pretty sure I have a rough idea how to do it. I'll probably start next week when I have some time.

Good idea! Can you, please, share your Anki cards via this forum, or personal message?