The New Method / Substep / Concept Idea Thread

[crafto22]

Speed-optimized is 11, and there's a reason it's used over optimal.

Also, seriously, what are your fingertricks for E and S slices?

I position my thumb and index finger and my thumb on the U layer, more specifically thumb on UF and index on UB. I use my middle finger to do E'. My other hand is holding the D layer with the thumb and middle finger and I use my index to to E. The same applies for S. It's pretty fast, and I have sub-oned this alg: M' E' M E' M' E2 M

 

[AlphaSheep]

It's pretty fast, and I have sub-oned this alg: M' E' M E' M' E2 M

Is it just me or can you save 3 moves by just doing M E' M' E?

 

[crafto22]

Is it just me or can you save 3 moves by just doing M E' M' E?

It doesn't matter, my point is I can do many slice moves quickly using my fingertrick.

 

[supercavitation]

I position my thumb and index finger and my thumb on the U layer, more specifically thumb on UF and index on UB. I use my middle finger to do E'. My other hand is holding the D layer with the thumb and middle finger and I use my index to to E. The same applies for S. It's pretty fast, and I have sub-oned this alg: M' E' M E' M' E2 M

So you regrip to do S moves (i.e., x' E' for S')? I'm trying to reach without one, and can't.

That seems to be the standard Roux finger trick for E-slices?

 

[PenguinsDontFly]

So you regrip to do S moves (i.e., x' E' for S')? I'm trying to reach without one, and can't.

That seems to be the standard Roux finger trick for E-slices?

we generally try to avoid E moves, but in LSE the most common alg using E moves is E2 M E2 M/M/M2. everyone has a different way of doing it. personally, I tilt the cube 45 degrees like a half z' and double flick E2 with right index and middle.

 

[supercavitation]

we generally try to avoid E moves, but in LSE the most common alg using E moves is E2 M E2 M/M/M2. everyone has a different way of doing it. personally, I tilt the cube 45 degrees like a half z' and double flick E2 with right index and middle.

That's pretty good. His finger trick above was what I used for dots when I was a roux user.

 

[shadowslice e]

Ok, so I posted this in the roux thread a while ago but I decided to post it here for those who don't go to that thread.

pre LSE edge control (case reduction)
1) Leave 1 misoriented and 1 oriented edge in D after F2B
2) If there are 3 misoriented edges in U, apply COLL algs
If there is one, execute a CLL+4Flip

Algs required:84

This is far less than if you didn't do case reduction for forcing LSE and gives you arguably the best case as you can do pairing easily, more easily even than perhaps a skip.

It is also easy to recognise; only fractionally harder than CMLL or OLL equivalent to maybe OLLCP (or slightly easier)

 

[crafto22]

Okay, so I though of this simply as a quick way to have a method with a decently lower move count then CFOP, yet by using similar concepts. When I note the movecount, I'm referring to someone who is pretty good at blockbuilding.

Steps

1) FB + FD (8 moves)
2) 2x2x2 in BDR (8 moves)
3) Winter Variation for last pair (8 moves)
4) OELLCP (9 moves)
5) EPLL (7 moves)

Movecount:
45 if you have okay blockbuilding
40 if you have pretty good blockbuilding
35? if you can find semi-optimal solutions during blockbuilding steps

 

[Praetorian]

Okay, so I though of this simply as a quick way to have a method with a decently lower move count then CFOP, yet by using similar concepts. When I note the movecount, I'm referring to someone who is pretty good at blockbuilding.

Steps

1) FB + FD (8 moves)
2) 2x2x2 in BDR (8 moves)
3) Winter Variation for last pair (8 moves)
4) OELLCP (9 moves)
5) EPLL (7 moves)

Movecount:
45 if you have okay blockbuilding
40 if you have pretty good blockbuilding
35? if you can find semi-optimal solutions during blockbuilding steps

whats OELLCP do you mean orienting edges whilst permuting corners, and how can it be done in 9 moves?

 

[obelisk477]

Okay, so I though of this simply as a quick way to have a method with a decently lower move count then CFOP, yet by using similar concepts. When I note the movecount, I'm referring to someone who is pretty good at blockbuilding.

Steps

1) FB + FD (8 moves)
2) 2x2x2 in BDR (8 moves)
3) Winter Variation for last pair (8 moves)
4) OELLCP (9 moves)
5) EPLL (7 moves)

Movecount:
45 if you have okay blockbuilding
40 if you have pretty good blockbuilding
35? if you can find semi-optimal solutions during blockbuilding steps

.....no.

WV is off by an extraordinary amount, because you have to form the pair first, and then the cases probably average ~9 moves after that. so maybe ~13 average moves, adding 5 to the total. Add a couple of AUFs in there and you're already up to Roux movecounts with worse lookahead

 

[supercavitation]

whats OELLCP do you mean orienting edges whilst permuting corners, and how can it be done in 9 moves?

If so, it's called CPEOLL, and based on Kirjava's algs, a basic (and probably very slightly off) calculation sets the average moves at 8.1 moves on average counting the skip case, and 8.27 without it.

EDIT: This is false, see below.

.....no.

WV is off by an extraordinary amount, because you have to form the pair first, and then the cases probably average ~9 moves after that. so maybe ~13 average moves, adding 5 to the total. Add a couple of AUFs in there and you're already up to Roux movecounts with worse lookahead

Winter Variation is an average of 8.07 moves, while F2L is an average of 6.7 moves a slot. Get rid of the insertion, and you've got 3.7 moves per pair, but add back a move for AUF, and you get 12.77 moves on average. Nice call.

This brings our estimate to 44 moves on average, ignoring AUF for both LL algs (so, say 45-46 on average). I'm also skeptical about your move count estimates for the first two steps, but I don't have data to back that up, so I'll use them for now. That's about what good Rouxers get (and you're assuming their level of block building efficiency for the 45-46), and as obelisk said, lookahead is worse.

EDIT: CPEOLL messes with Corner Orientation, so your move count would actually be higher than I predicted

 

[shadowslice e]

Okay, so I though of this simply as a quick way to have a method with a decently lower move count then CFOP, yet by using similar concepts. When I note the movecount, I'm referring to someone who is pretty good at blockbuilding.

Steps

1) FB + FD (8 moves)
2) 2x2x2 in BDR (8 moves)
3) Winter Variation for last pair (8 moves)
4) OELLCP (9 moves)
5) EPLL (7 moves)

Movecount:
45 if you have okay blockbuilding
40 if you have pretty good blockbuilding
35? if you can find semi-optimal solutions during blockbuilding steps

https://www.speedsolving.com/forum/...Immune-System-(solving-method-potential-1LLL)

 

[AlexMaass]

Is it just me or can you save 3 moves by just doing M E' M' E?

http://niceme.me/

 

[supercavitation]

EDIT: CPEOLL messes with Corner Orientation, so your move count would actually be higher than I predicted

I took a look at some OLLCPs that don't mess with Corner Orientation, but do permute Corners and orient Edges (20,28, and 57, and PLLs for EO skips), and came up with an average of 10.15 moves counting the skip case and 10.35 not counting it. So to redo my math from last night (accepting your move counts for blockbuilding for lack of a better option):

First Block+FD: 8
2x2x2 in BDR: 8
Winter Variation for the last pair: 12.77
EOLLCP: 10.15
EPLL: 6.75 using the usual MU algs

Total Movecount: 45.67

Add in 1 or 2 moves for AUF in LL, and move count becomes almost identical to Roux.

 

[crafto22]

I took a look at some OLLCPs that don't mess with Corner Orientation, but do permute Corners and orient Edges (20,28, and 57, and PLLs for EO skips), and came up with an average of 10.15 moves counting the skip case and 10.35 not counting it. So to redo my math from last night (accepting your move counts for blockbuilding for lack of a better option):

First Block+FD: 8
2x2x2 in BDR: 8
Winter Variation for the last pair: 12.77
EOLLCP: 10.15
EPLL: 6.75 using the usual MU algs

Total Movecount: 45.67

Add in 1 or 2 moves for AUF in LL, and move count becomes almost identical to Roux.

You're right, I forgot to factor pair building. EPLL is not 6.75 with the MU algs, its 7.7
OELLCP preserves CO and it's 9 moves average using the algs on Sarah's Cubing Site.

So average moves would actually be more like 43 for people with decent blockbuilding. Also, M-CELL is not the same as what I am suggesting. Although it is similar, my method is not a 1LLL and seems to require less algs. The steps and different in multiple ways, but class just started, so I can't keep writing.

 

[shadowslice e]

You're right, I forgot to factor pair building. EPLL is not 6.75 with the MU algs, its 7.7
OELLCP preserves CO and it's 9 moves average using the algs on Sarah's Cubing Site.

So average moves would actually be more like 43 for people with decent blockbuilding. Also, M-CELL is not the same as what I am suggesting. Although it is similar, my method is not a 1LLL and seems to require less algs. The steps and different in multiple ways, but class just started, so I can't keep writing.

Well, if you can do FB+DBR (which incidentally I call SSF2L) in <16 moves in a speedsolve I would advise you to take up roux or FreeFOP or petrus. I would say it's more like
FB+DF-10
2x2x2- 8
Pair-4
WV-8
OELLCP- 9
ELL: 8

I rounded but the average is still ~47.

All in all it has the potential to be a good method but I think you're being a bit too idealistic on the block building movecounts in a speedsolve.

 

[supercavitation]

You're right, I forgot to factor pair building. EPLL is not 6.75 with the MU algs, its 7.7
OELLCP preserves CO and it's 9 moves average using the algs on Sarah's Cubing Site.

So average moves would actually be more like 43 for people with decent blockbuilding. Also, M-CELL is not the same as what I am suggesting. Although it is similar, my method is not a 1LLL and seems to require less algs. The steps and different in multiple ways, but class just started, so I can't keep writing.

M-CELL wasn't my suggestion, not sure who that was to.

Which EPLLs did you use when calculating that? In STM, if you account for the likelihood of various cases, it's 6.75 (7 for both U perms and H perm, 9 for Z perm). Accounting for probability, its (7*4+7*4+7*1+9*2)/12=6.75.

Also, with Sarah's Algs, if you account for probability (and the EO skip cases), you're still at 9.90 moves, so that saves you 0.25 moves.

 

[shadowslice e]

M-CELL wasn't my suggestion, not sure who that was to.

Which EPLLs did you use when calculating that? In STM, if you account for the likelihood of various cases, it's 6.75 (7 for both U perms and H perm, 9 for Z perm). Accounting for probability, its (7*4+7*4+7*1+9*2)/12=6.75.

Also, with Sarah's Algs, if you account for probability (and the EO skip cases), you're still at 9.90 moves, so that saves you 0.25 moves.

M-CELL was my suggestion.

The movecount is still decent to be honest at 10+8+4+8.7+9.9+7.5= 48.1 movesish

Also, I'm interested in how you managed to get a lower average movecount than the best case in your EPLL set...

 

[supercavitation]

M-CELL was my suggestion.

The movecount is still decent to be honest at 10+8+4+8.7+9.9+7.5= 48.1 movesish

Also, I'm interested in how you managed to get a lower average movecount than the best case in your EPLL set...

Skip case occurs 1/12. So I guess that should really read (7*4+7*4+7*1+9*2+0*1)/12=6.75. Also, you forgot to account for AUF in your move count total.

Also, WV is 8.07, not 8.7.

 

[crafto22]

You are right, I forgot to account for skips. Still a skilled user could surely achieve 43ish moves somewhat consistently. I think some modifications to this could make it good!