The New Method / Substep / Concept Idea Thread

[Teoidus]

Better idea:

CSP
F2B
CLL + DFDB (at max 3 * 12 = 36 algs)
EPLL

Not a method idea, but a hardware idea: is it illegal to attach little rubbers to the corners of your cube? I feel like that would give great advantage for table abuse OH (e.g. LSE)

 

[Neuro]

I went through and did 12 speedsolves with ZZ-Tri, and my mean was about 48.92. However, I did have 2 solves that would be considered rare (both were in the 30's.) I still consider the average movecount to be between 50-55 moves, as 2/3rds of the solves fell into this range and is captured by a 95% confidence interval (38.45 - 59.38.) So I believe I was correct in my assumption that ZZ-Tri is in the same move range as ZZ-CT.

 

[Teoidus]

Assuming a normal distribution, a 95% confidence interval (38.45, 59.38) gives an expected a mean movecount of (38.45 + 59.38) / 2 = 48.915. This is very close to your avg12 = 48.92.

Though it seems likely your distribution is more left-skewed, I would still be interested to see how avg movecount turns out given a larger sample size.

I'm tempted to think that NLS actually reduces to quite a friendly ZBLL subset (since there's a 1/2 * 2/3 = 1/3 chance of getting a pure commutator). You don't get pure sunes and such, but you get A9s, Niklases, sexysldg variants, etc.

Of course at the same time you get pure twists, so. It's hard to say.

 

[Neuro]

What sample size would you reccomend? Simply go to n=30?

 

[Shiv3r]

Not a method idea, but a hardware idea: is it illegal to attach little rubbers to the corners of your cube? I feel like that would give great advantage for table abuse OH (e.g. LSE)

It might be if you make the rubber pieces the stickers themselves, and then color them. then they'd be like tiles, I think, to the WCA.

 

[Y2k1]

https://docs.google.com/document/d/1kKIMOr6LSj7F_uxzmd2N077oc53lhQ3Ki0P6oO_Ojzs/edit?usp=drivesdk

For lin method (not mine)

Edit: changed link

 

[Shiv3r]

https://docs.google.com/document/d/1aTIOxIbhA6UdQqAGE3QRYErBsRNbt2g5DB5Ag4iXMv8/edit?usp=drivesdk

For lin method (not mine)

it doesn't let me access it. I want to look at it though, I like the Lin method a lot.

 

[Y2k1]

it doesn't let me access it. I want to look at it though, I like the Lin method a lot.

https://docs.google.com/document/d/1kKIMOr6LSj7F_uxzmd2N077oc53lhQ3Ki0P6oO_Ojzs/edit?usp=drivesdk

 

[Shiv3r]

https://docs.google.com/document/d/1kKIMOr6LSj7F_uxzmd2N077oc53lhQ3Ki0P6oO_Ojzs/edit?usp=drivesdk

thank you, this is totally worth it. who made this?
Also I am just using the algs that were written in the description, they seem ok.

 

[Y2k1]

thank you, this is totally worth it. who made this?
Also I am just using the algs that were written in the description, they seem ok.

I actually heard about the method myself a couple months ago, and someone shared that doc when I was looking for the algs, but I dont remember who

Edit: it was Cale S

 

[Shiv3r]

I actually heard about the method myself a couple months ago, and someone shared that doc when I was looking for the algs, but I dont remember who

Edit: it was Cale S

Nice. I found a video describing it made in january, and so I looked at that and the algs.

 

[efattah]

Expanded 31-page LMCF description now includes FULL modernized & optimized Waterman L6E, which I believe is the first true documentation that difficult last step. Of interest is that using Waterman L6E a Roux solver could theoretically leave any edge in the second block disoriented and it would be solved during the LSE phase.

https://drive.google.com/open?id=0B2QnZ3uD6I8kUmM0V2RTdWg1U2M

 

[Teoidus]

What sample size would you reccomend? Simply go to n=30?

Yeah 30 or 40 would be nice

 

[Calode]

I took some time to get started a ZZ-tripod spreadsheet that comes with NLS sheet and an excerpt of tripod zbll so people who don't know zbll (me) don't have to go scavenging for algs.

here: https://docs.google.com/spreadsheets/d/1uZkgPNUNwMcyMN0xJtV8m7bxzhr1fQ1l9pYmdlzTeSk/edit?usp=sharing

@Neuro
I notice numerous mistakes in NLS that you mentioned. I'm more than happy to help gen NLS stuff. PM me your email and I can add you as a contributor.

 

[Neuro]

I took some time to get started a ZZ-tripod spreadsheet that comes with NLS sheet and an excerpt of tripod zbll so people who don't know zbll (me) don't have to go scavenging for algs.

here: https://docs.google.com/spreadsheets/d/1uZkgPNUNwMcyMN0xJtV8m7bxzhr1fQ1l9pYmdlzTeSk/edit?usp=sharing

@Neuro
I notice numerous mistakes in NLS that you mentioned. I'm more than happy to help gen NLS stuff. PM me your email and I can add you as a contributor.

Thanks for the sheet, it should help a lot! Use [email protected] to contact. Unfortunately I don't have much experience in making "good" algs (simply use cube explorer and look for good cases), so this should help progress ZZ-Tri significantly.

 

[Shiv3r]

JAcube is better than CE imo.
Also, I have looked a little more at the Lin method. As a lot of people use Roux-n-screw for its lack of a lot of algs, I can see 2 ways to get fast at squan: go Vandenbergh, and learn algs all the way, or go R&S->RLBL(what Helmer Ewert uses: Roux F2B and then insert DF&DB and then LL) ->Lin->Lin w/CSP, or even Lin algs where you insert DF edge and solve parity at the same time might be nice.
I think that Lin w/ CSP has the possibility to be a major contender to Vandenbergh even with the low amount of algs relative to vandenbergh.
Im halfway done with Learning DF+CLL algs, and my times have already improved.

 

[Y2k1]

Making another set that solves parity while inserting DB is a great idea, where are squan algs genned?

 

[Shiv3r]

Making another set that solves parity while inserting DB is a great idea, where are squan algs genned?

https://www.jaapsch.net/puzzles/square1.htm, there's an optimal squan solver you can download. I don't believe it allows you to ignore pieces however. I don't own any windows computers(its a .exe) so I will not be able to be of very much help. If you want to gen the algs though, be my guest.

 

[efattah]

Here is LMCF revision 3, expanded to 38 pages, includes discussion on the beginner method and more discussion on advanced concepts and L6E comparison solves at the end:
https://drive.google.com/open?id=0B2QnZ3uD6I8kZ0NvaDM5SlphTjg

Here is a link to training sets & images you can run through any slide show to practice algorithms, or print out some of the sheets:
https://drive.google.com/open?id=0B2QnZ3uD6I8kdHNFT2FGcXRUaFE
https://drive.google.com/open?id=0B2QnZ3uD6I8kdHNFT2FGcXRUaFE

 

[Shiv3r]

What's the total alg count for NLS+TELL(Tripod LL, Edges oriented, another name for Tripod ZBLL), and how does it compare to say COLL+EPLL, or phasing+ZZLL?

One thing: This is glorious. I did a solve, and Its muuuch better than the other LS methods, because you don't waste moves setting up the pair. my first solve was CE alg into a Pure flip which I already knew the alg for. Good thing about this, is that I probably know about 50% of the algs already becuz I can BS most of L3C with comms, and I know the pure flips, and PLL.

EDIT: did another solve, got a PLL. god, when will I ever get a ZBLL?