The New Method / Substep / Concept Idea Thread

[Abram Lookadoo]

Second, you're using a giant algset to force a tiny, bad set of cases

the reason I'm using this set off ll algs is that, for the first set of ll-eo, you aren't held back from the edges orientation, the algorithm will be better and more efficient the advantage for the second step of eoll is that there is a 1/8 skip chance. i was afraid that the 2x2x3 block would be a bit hard to conceive in 15 seconds, but if enough practice went into seeing the block and how pieces intertwine, we should be able to find a way to make this happen.

 

[PapaSmurf]

new method (BFLE)

1) create a 2x2x3 block in DB
this step should be seen in inspection (9-15 moves)

2) solve the 1x2x3 block in FD
solve the block using intuition (13-21 moves)

3) LL-eo
use a speed optimal algorithm that solves everything in the U layer ignoring eo (1/496 algs)(7-15? moves)

4) eo
solve eo using a speed optimal algorithm (1/8 skip chance)(1/3 algs)(0-14 moves)(11.25 moves avg)

So this is basically freefop with 493 (I think, because same as zbll) algs then eo. You’re better off doing oll/pll, as pure edge flip algs aren’t very good. I guess you get good skip chances, but vanilla cfop is better. As is vanilla roux or zz.

 

[Abram Lookadoo]

You’re better off doing oll/pll, as pure edge flip algs aren’t very good.

i know pure edge flips are not that good. this is what caused me to think of ignoring the eo until unavoidable. this allows for better algs for LL-eo, so if you get a skip, it will be better than using zbll when edges are already oriented. i could be wrong on this, but i always thought zbll was 496 algs. am i wrong?

 

[PapaSmurf]

i know pure edge flips are not that good. this is what caused me to think of ignoring the eo until unavoidable. this allows for better algs for LL-eo, so if you get a skip, it will be better than using zbll when edges are already oriented.

I don’t see the logic in this tbh, as FRUR’U’F’ is a lot better than doing the pure edge flip, and the chance of a skip is exactly the same. And zbll algs are already developed and fast, so I do see why without caring for eo the algs could be better, but it won’t be significant.

i could be wrong on this, but i always thought zbll was 496 algs. am i wrong?

I just checked the wiki, it says 493.

 

[Abram Lookadoo]

I don’t see the logic in this tbh, as FRUR’U’F’ is a lot better than doing the pure edge flip, and the chance of a skip is exactly the same. And zbll algs are already developed and fast, so I do see why without caring for eo the algs could be better, but it won’t be significant

there are some algorithms (≈7/8) that are faster if you ignore edge orientation. the reason i prefer LL-eo → eo is that nearly all solving is done is when there is less to mess up. and to counter the fact that zblls are more developed here is a website with every single ll case (17 moves or less) and how fast it is. (http://birdflu.lar5.com/?pos=____&list=algs)

 

[Teoidus]

i know pure edge flips are not that good. this is what caused me to think of ignoring the eo until unavoidable. this allows for better algs for LL-eo, so if you get a skip, it will be better than using zbll when edges are already oriented. i could be wrong on this, but i always thought zbll was 496 algs. am i wrong?

> I know the pure edge flips aren't good
> Let's propose a method with pure edge flips!

?????????

 

[1001010101001]

New 4x4x4 Reduction method( this might have been thought of before, but don't think I am copying as I thought of it myself)
1. 2 opposite centers
2. Pair 6 edges and put them on the R and L slice ( 2 or 3 moves each)
3. Remaining centers
4.Pair remaining edges using 6- reduction method
5.3x3 stage
Basically I modified centers first and had a more efficient edge pairing.

 

[xyzzy]

3. Remaining centers

This step sucks because it's extremely heavy on slice moves.

Edge pairing after centres are done takes ~3.5 moves per dedge, compared to ~3 moves per dedge if you do it between the first two centres and the other centres, which means that you're saving (3.5−3)×6 = 3 moves on edge pairing compared to plain reduction, and saving 3 moves doesn't make up for how you need to use slice moves throughout the centres.

(I like the idea though, and I've experimented with keeping oriented edges on L and R to make transition into EOline easier, but the centres are just impossible to do quickly.)

 

[Thom S.]

new method (BFLE)

1) create a 2x2x3 block in DB
this step should be seen in inspection (9-15 moves)

2) solve the 1x2x3 block in FD
solve the block using intuition (13-21 moves)

3) LL-eo
use a speed optimal algorithm that solves everything in the U layer ignoring eo (1/496 algs)(7-15? moves)

4) eo
solve eo using a speed optimal algorithm (1/8 skip chance)(1/3 algs)(0-14 moves)(11.25 moves avg)

Not saying that this method is good, but try your EO with FR and/or FD flipped, as they may or may not have better pure flip Algorithms

 

[Abram Lookadoo]

> I know the pure edge flips aren't good
> Let's propose a method with pure edge flips!

?????????

the idea here is that edge orientation could still be a problem when other pieces are unsolved, and if edges are solved finger tricks are less used. so instead this method hopes for the 1/8 chance of getting a skip on the last step.

 

[PapaSmurf]

the idea here is that edge orientation could still be a problem when other pieces are unsolved, and if edges are solved finger tricks are less used. so instead this method hopes for the 1/8 chance of getting a skip on the last step.

So this is a method built for singles. 1/8 sounds high, but recognition would be bad for the edges, and you’re more likely to not get a skip in an ao 5 than you are, then you have bad flip algs to do. CFCE is probably better, and CFOP is definitely better than CFCE. (And roux and zz are better than CFOP)

 

[Abram Lookadoo]

So this is a method built for singles. 1/8 sounds high, but recognition would be bad for the edges, and you’re more likely to not get a skip in an ao 5 than you are, then you have bad flip algs to do. CFCE is probably better, and CFOP is definitely better than CFCE

yes, this method was made with singles in mind.

is it really proven that cfop is better than cfce?

 

[Teoidus]

1/8 skip chance means nothing without context

Here's a method I call (new Array(4).fill(0).map(e => String.fromCharCode(32 + 95 * Math.random())).join("")):

1. Solve everything but the last corner
2. Solve the last corner

100% SKIP CHANCE GUYS!!!!!!! THE LAST STEP IS GUARANTEED TO SKIP.

OMG. THIS IS A METHOD MADE FOR SINGLES

 

[Abram Lookadoo]

1/8 skip chance means nothing without context

Here's a method I call (new Array(4).fill(String.fromCharCode(32 + 95 * Math.random()).join("")):

1. Solve everything but the last corner
2. Solve the last corner

100% SKIP CHANCE GUYS!!!!!!! THE LAST STEP IS GUARANTEED TO SKIP.

OMG. THIS IS A METHOD MADE FOR SINGLES

i did give the 1/8 chance context, as an eo skip.

your method is really bad.
also the name is hard to pronounce.

 

[Teoidus]

But why have 7/8 * pure edge flip movecount + 1/8 * 0 when you can have like, on average, 5 in normal petrus?

If you want 7/8 * pure edge flip movecount to be on average the same or better than 5,
the average movecount of pure edge flip algorithms (with AUF) has to be at most 40 / 7 or 5.71428...

Do you really think LL pure edge flip algorithms average 5.7 moves?

Or, more succintly put
you could just use petrus

 

[PapaSmurf]

yes, this method was made with singles in mind.

is it really proven that cfop is better than cfce?

Firstly, there's no point for methods with singles in mind. It'll never be used at a high level. Also, try CFCE for a month, then CFOP. Which one is better? CFOP - better recognition, better algs. Simple. And I agree with @Teoidus . Petrus is better. And the point of his bad method was to illustrate that skip methods aren't good. Especially if you depend on skips.

 

[Teoidus]

The simple calculation with petrus EO does not even take into account the astronomical amount of effort you spend forcing a pure flip ELL. In fact, with that same effort, I'm pretty sure you can just solve the LL in one alg...

 

[Abram Lookadoo]

In fact, with that same effort, I'm pretty sure you can just solve the LL in one alg...

it would take more effort because the algorithms are longer.

if you ignore eo, it will almost always result in a better algorithm, and if paired with a skip it will be faster than if you solve eo then use one alg, or even get an eo skip in petrus, and finish with zbll.

 

[Teoidus]

Yes, but notice you already lost on average 5 moves vs petrus's EO to do your pure flips. (pure flip ELL generously averages 10 moves)

So unless you are saying that your algorithms are so good that they average less than 9 moves (ZBLL averages around 14), I would be very surprised if your method is any more efficient than petrus.

 

[Abram Lookadoo]

average 9 moves

out of 12 tests the average for speed optimal algs is 12.25 htm
so it is more efficient if you get a skip for eo