Method similar to ZB

[Davepencilguin]

I'm tentative about posting this, but I haven't found anything in the archives, so I might as well bring it up.

The idea of ZB is to reduce the last layer possibilities down to a "reasonable" level by orienting the edges first.
My question is why not orient the corners instead to reduce that number down farther?

I understand that there would be a much higher number of algorithms needed for this complete F2L set than for ZBF2L, but (at least for the beginning of developing) a solver could stick to the winter variation.

As for the last layer, it comes down to the 21 permutations x 8 edge orientations, making 168 algorithms needed, which isn't that bad compared to ZBLL.


Ideas/comments?

 

[hawkmp4]

So you're saying, do cross + 3*f2l slots + (name of your method's last slot) + PLL?
You're essentially combining the last slot with OLL. Bad idea, I think.
Do you have any algorithms or example solves?

 

[Weston]

Sticking to winter variation is kind of like doing VH+ZBLL.
If youre going to learn ZBLL, you should go all the way and learn ZB.
I don't know much about statistics (I'm in calculus) but since there are more F2L cases (42)than there are PLL cases (21), I think it is less algs to do corner orientation with PLL than corner orienatin with ZBLL.

There are 27 winter variation cases, so I'm guessing that there are that many cases for each f2L case, since WV is just a subset of "this method."
so 42*27=1134 cases for the last slot.
(I might be completely off)
Then for the LL there are the 168 cases you mentioned. Probably actually less though.

Thats quite a bit more algs than ZB.
That totals 1302 non distinct cases.

 

[Weston]

So you're saying, do cross + 3*f2l slots + (name of your method's last slot) + PLL?
You're essentially combining the last slot with OLL. Bad idea, I think.
Do you have any algorithms or example solves?

No thats not what he meant. That would be around 42*27*8=9072 cases.
Minus the doubles obviously.

 

[hawkmp4]

Ah, okay. I've got it now. Sorry, it's 2:30 local time. Very tired.
I'm really too tired to think about this but I'd like to see the number of algorithms for this. I think that's the most important part.

 

[rubiknewbie]

There were some related discussion here:
http://www.speedsolving.com/forum/showthread.php?t=8820

 

[eastamazonantidote]

I actually had this very thought about 6 months ago and spent about a week thinking about it. Though feasibly an alternative to ZBLL, the lake of 2GLLs makes it slightly less attractive despite the smaller number of cases. However, <M,U,R> gen cases are nice. hmmm.....



EDIT: Misunderstanding at first. Changed post entirely.

 

[Cyrus C.]

I think it's quite a good idea, I don't understand what's wrong with just using winter variation.

 

[Weston]

I think it's quite a good idea, I don't understand what's wrong with just using winter variation.

Because the idea of ZB is to reduce the number of steps. Using winter variation will go something like.
Using winter variation essentially adds another step because you can't make an f2l pair and insert it right away. You have to AUF and pause to recognize the case, then do the alg, then LL. Thats the same number of steps as Fridrich.

 

[masterofthebass]

This idea is sort of like doing CLS without ELS. Pretty pointless IMO, since ELS takes almost no time, and just leaves you with PLL instead of this idea.

 

[Davepencilguin]

Ok, so now that I think about it, there are actually a HUGE number of xxF2L cases.
But if you set up the cube so that neither the F2L corner nor the F2L edge are in the slot and so that the corner is not oriented so the cross color is facing up, the number of xxF2L algorithms would be (after AUF) 27 x 4 x 2 = 216 algorithms + mirrors

The problem with this is the cube setup to get one of these 216 cases. But these cases wouldn't be THAT bad because the case would be one of the 216 algorithms + either a RUR' or a RU2R' added to the beginning.
As for the cases with the corner oriented incorrectly, proper control during the F2L solve would eliminate the problem.
(I haven't taken a statistics class, so I'm not going to attempt to find out the chances of the corner/edge piece being in the slot.)

 

[Tim Reynolds]

Ok, so now that I think about it, there are actually a HUGE number of xxF2L cases.
But if you set up the cube so that neither the F2L corner nor the F2L edge are in the slot and so that the corner is not oriented so the cross color is facing up, the number of xxF2L algorithms would be (after AUF) 27 x 4 x 2 = 216 algorithms + mirrors

The problem with this is the cube setup to get one of these 216 cases. But these cases wouldn't be THAT bad because the case would be one of the 216 algorithms + either a RUR' or a RU2R' added to the beginning.
As for the cases with the corner oriented incorrectly, proper control during the F2L solve would eliminate the problem.
(I haven't taken a statistics class, so I'm not going to attempt to find out the chances of the corner/edge piece being in the slot.)

Corner control is way harder than edge control. What's easy about edge control is that, if you use 2-gen stuff, edges don't flip. Corners do twist though.

As for the probability you mention, it would be roughly 1-(4/5)(4/5)=36%. So about a third of the time, you're going to have to do a set-up to get the pieces out of the slot; it's about half the time if you include the mis-oriented corner case. After recognizing that you have to do such a set-up move, you then have to recognize which case you have, which I would imagine would be more difficult then ZBF2L recognition.

Unless you could do all of xxF2L in one alg, I don't think it would be a good idea. The only other option I can see is a VHF2L style thing--first make a pair, then use Winter Variation algs to put it in while orienting corners. But I don't think VHF2L is that great a method, so I don't particularly like this idea.