ZZ-CT Thread

[alwin5b]

I think a big appeal of Chris's algs aren't that they're necessrily the most efficient, but that they're 2 gen, which makes them all insanely ergonomic.

I agree. Given the choice between two equally fast algs, I always choose the one with fewer moves (in any metric). That's just a personal preference of mine. Most of the new algs I found are <R, U, L> which is slightly worse than <R, U>, but since they are shorter it evens out. (also, I think its cool to complement the ZZ method, which was designed to solve the cube with <R,U,L> moves, with <R,U,L>-algs. But thats again personal preference)

I mistakenly wrote that my algs are faster, what I meant was that some of the algs are faster and most are about equally fast, but none are slower.

If you're not already doing so, limit the moves you search. I generate algs in <RUF> <RUD> and <RUL> first, then generate <RUFD>, and <RUDL> <RUFB> algs (these algs are often nice with rotations/wide moves added in).

I generated first unlimited, then <R,U,L> if I got too many good algs; both for face turn metric and quarter turn metric. I will definetely use your strategy for the longer algs, thanks.

 

[JTcuber]

Hey, does anyone know the approximate alg count if you were to look at CP in TSLE, so you could force a 2 gen case?(Right opp or front opp)

 

[4Chan]

Hey, does anyone know the approximate alg count if you were to look at CP in TSLE, so you could force a 2 gen case?(Right opp or front opp)

It's around 650

 

[MethodNeutral]

I think this method is going great places, but I really don't want to spend so much time learning TSLE... Maybe over the summer, I'll see. I keep seeing you all saying a bunch of cases are easy to reduce, but there are just so many... Anyway, my way of doing it for about 3 months now is just to insert the edge by itself during F2L, then I use CLS (I or Im) on the corner that happens to be with it in the slot. If it happens to be oriented correctly, I just use an OcLL. It's really nice, because if you're doing F2L and you don't see anything you can just insert the edge while looking for a block and then get back to it later. Any thoughts??

 

[JTcuber]

It's around 650

Damn. There goes that idea.

 

[Teoidus]

You could force CP during left block instead, porky style

i think it just ends up adding moves tho

 

[MethodNeutral]

You could force CP during left block instead, porky style

i think it just ends up adding moves tho

If you do it after left block it's at most 5 moves, but recognition is always there.

 

[Teoidus]

Well it's still 5 moves on top of TSLE and TTLL, and I'm pretty sure 2gen TTLLs aren't any shorter than normal TTLLs

so the trade off is less algs and 2gen algs, but more moves and the recog as you said

Not sure if it's worth it

 

[MethodNeutral]

Well it's still 5 moves on top of TSLE and TTLL, and I'm pretty sure 2gen TTLLs aren't any shorter than normal TTLLs

so the trade off is less algs and 2gen algs, but more moves and the recog as you said

Not sure if it's worth it

Yeah, I'd just stick to regular TTLL's. Besides, if you like mirroring, you can learn the subsets with headlights in front and opposites in front and mirror all of those so that you know 2/3 of TTLL. That's what I'm doing so far.

Also, instead of TSLE I've been doing a CLS approach where I just insert the edge then use I or Im CLS on whatever corner is in the slot with it. Any comments?

 

[Teoidus]

Well it's less algs and higher skip chance but more moves and more looks

Might be good as a stepping stone to ZZ-CT, but if you're thinking if making that your primary way of solving LS+LL it's probably better to just do LS then COLL/EPLL

 

[MethodNeutral]

Well it's less algs and higher skip chance but more moves and more looks

Why is it more looks? Instead of TSLE+TTLL I do CLS+TTLL.

 

[MethodNeutral]

Also, I just realized that while I'm doing this, I could just insert the edge then learn the TSLE cases for when the edge is inserted. That's way more achievable for me, and keep a relatively low movecount compared to my CLS/OcLL approach.

 

[JTcuber]

Also, I just realized that while I'm doing this, I could just insert the edge then learn the TSLE cases for when the edge is inserted. That's way more achievable for me, and keep a relatively low movecount compared to my CLS/OcLL approach.

Yeah, that's what I've done as a stepping stone to full TSLE. It's only 20ish algs

 

[lillod]

And is anyone else having problems win the tsle anki file?

Yes the tsle deck doesn't seem to import anything. The ttll deck works fine for me and I tagged the cards by corner permutation. I'm planning to use it to learn a set of ttll a week.

 

[Allahjabark]

Hey Chris, I just found out about your method, as I just started cubing about a month ago. I know about half of winter variation and am wondering if it can replace tlse if I use a pretend corner.

 

[AlphaSheep]

Hey Chris, I just found out about your method, as I just started cubing about a month ago. I know about half of winter variation and am wondering if it can replace tlse if I use a pretend corner.

Yes it can. All of the really short winter variation cases (7 moves or less) make really good TSLE algs and I think most people use those. There are several other cases where a decent strategy is to do a single trigger (e.g. R U R') to reduce to a pseudo-winter variation case.

There are some bad winter variation cases though where it is not worth it. When picking TSLE algs, it helps to remember that every single TSLE case can be solved in 11 moves or less, even using only R and U moves.

 

[lillod]

So when you scramble the cube <R,U> you will always get the front opposite 2 gen set of ttll when you solve it. I thought if you did something like D' R U R' D and then scrambled the cube 2 gen it would constrict the ttll to another set. Instead it seems that it can be any of the non 2 gen ttll. Why is this?

 

[MethodNeutral]

Basically, it all depends on how the cube is scrambled before doing the moves and which pair you want to leave for TTLL. I'm assuming you always solve the back pair, then use the front right slot for TTLL. If you start from a solved state, you will get a case with headlights on the right. This is because the moves you do (D' R U R' D) swap the two corners at URF and URB on a solved cube. If you do a U2 before these moves and try a TTLL, you'll get a double headlights case because this time the ULF and ULB corners are swapped. A different AUF would give a different case because different corners are being swapped. I hope this helps, I know that was hard to follow.


Sent from my iPhone using Tapatalk

 

[xyzzy]

So when you scramble the cube <R,U> you will always get the front opposite 2 gen set of ttll when you solve it. I thought if you did something like D' R U R' D and then scrambled the cube 2 gen it would constrict the ttll to another set. Instead it seems that it can be any of the non 2 gen ttll. Why is this?

Assuming you have the cube in scramble orientation (adjust the colours accordingly if you use another orientation for ZZ), the D' R U R' D effectively swaps the white-red-green and white-blue-red corners. Let's say you have a 2-gen TTLL (not counting PLL, for simplicity); where can the white-red-green and white-blue-red corners be? There are essentially four possible locations for these two corners, so you can get four out of five of the other TTLL sets by swapping those two. (Note that there's still a sixth set which you will not encounter with a D' R U R' D setup, which I think is headlights-on-front.)

(also, ninja'd)

 

[lillod]

Basically, it all depends on how the cube is scrambled before doing the moves and which pair you want to leave for TTLL. I'm assuming you always solve the back pair, then use the front right slot for TTLL. If you start from a solved state, you will get a case with headlights on the right. This is because the moves you do (D' R U R' D) swap the two corners at URF and URB on a solved cube. If you do a U2 before these moves and try a TTLL, you'll get a double headlights case because this time the ULF and ULB corners are swapped. A different AUF would give a different case because different corners are being swapped. I hope this helps, I know that was hard to follow.


Sent from my iPhone using Tapatalk

What I was trying to say was that if you did the alg from the solved state you wont always get a ttll with headlights on right.