SZ method for 3x3 speed cubing , reducing zbll from 493 to 93

I-am-nishan

Regular
Joined
May 11, 2026
Messages
49
Location
My house
SZ method

Steps:

1. SZ-cross + Pair

solve the left and right and back white bottom edges plus one f2l pair, pair is optional for beginners and intermediates, this frees m' slice for the rest of the solve(M also possible but M' recommended)

2. SZ-F2L

remaining pairs with m slices, this allows to do f2l rotationlessly and efficiently, Only 4 M moves needed maximum

3. EOFLL

Orient last 5 edges, 4 yellow + remaining white edge, can be done intuitively or algorithmically

4. SZBLL

main original thing, 6 algs based on sune/anti sune that insert the last white edge AND force t case of zbll at same time, its deterministic not rng trust me

example: U R U R' U R U2 R' U M' U2 M

5. T-ZBLL+PLL(lucky chance for pll)

72 t subset zbll cases

total 104 algs with szbll and eofll

Varients

SZ-CMLL, instead of EOFLL+ u do cmll and then do roux style eo, and at last insert the last edge and do EPLL(4 algorithms), or do L5E

SZ-OP - instead of szbll u insert the white edge and do ocll (only 7 cases) and pll(21)

ALGS ON HERE

Spreadsheet best on dark mode

Join our discord server!
 
Last edited:
Maybe not, a niche method DOES NOT deserve that much attention
1000001559.webp
 
So uhh I have a few stuff about SZ which I haven't said(I wanted to prove) f2l only needs 4 M moves at maximum if you are efficient done a lot of solves and I use the m slice to oriented some e layer edges to avoid f moves, and second its okay if the front cross piece get solved somehow during f2l or(more commonly) eofll ,as they don't affect szbll anyway and its okay to break the front cross piece too (while doing f2l) , most of you guys already probably knew this but just saying for the people who don't, and also we got a discord server for absoulely no reason for sz
 
This is a nearly identical method to one posted a few days ago. The only difference is instead of inserting the cross edge during EO and simply being left with ZBLL, you do EO first, then SZBLL, then a subset of ZBLL. The problem is, as I said in that other thread, solving something incorrectly and then fixing it later is usually less efficient than just solving it correctly in the first place. In this case, solving EO, then doing SZBLL followed by the ZBLL subset, is just going to be slower than solving EO and DF at the same time, followed by full ZBLL. There may be fewer algorithms to learn with your idea, but you would want to transition to full ZBLL in the long run anyway.

The bigger problem, unfortunately, is that simply solving the DF edge alongside the rest of the cross is just going to be faster than solving it in a different way later, but then you're just left with regular CFOP.

As I said in that other thread, though, this doesn't mean you should give up on method creation. We all start somewhere, and it's a really cool and worthwhile art. But it does pay to think critically about your method before posting, making sure it's not only unique, but it holds some kind of advantage over existing methods. When you come up with something else, feel free to post it in the existing method ideas thread: https://www.speedsolving.com/threads/the-new-method-substep-concept-idea-thread.40975/
 
Back
Top