[Help Thread] ZZF2L Efficiency Tips? + Possible Roux SB Tips

[ZZeroingZZemdegs]

I really need tips on ZZF2L efficiency, although most tutorials don't have any tips on ZZF2L efficiency and look-ahead. Please keep look-ahead discussions to a minimum, as I already know that a great look-ahead leads to a more efficient ZZF2L. What I need more is tips on creating 3x2x1 blocks in the least moves possible, as my ZZF2L ranges from 20-30 moves, and I want to get that down to <20 moves. Anyone have any suggestions?

 

[Sue Doenim]

http://cube.crider.co.uk/zz.php?p=f2l

 

[ZZeroingZZemdegs]

Yes, I have looked at Conrad Rider's tutorial, but his tutorial for ZZF2L shows how to do F2L by making 2x2x1 blocks and then making 3x2x1 blocks by adding a 1x2x1 block. What I want to do is to create a 3x2x1 block in the least moves possible and use a "Roux-ish" block-building approach. Does anyone have tips on Roux SB that I could use for ZZF2L?

 

[Sue Doenim]

Yes, I have looked at Conrad Rider's tutorial, but his tutorial for ZZF2L shows how to do F2L by making 2x2x1 blocks and then making 3x2x1 blocks by adding a 1x2x1 block. What I want to do is to create a 3x2x1 block in the least moves possible and use a "Roux-ish" block-building approach. Does anyone have tips on Roux SB that I could use for ZZF2L?

Being a [somewhat mediocre] Roux and ZZ solver myself, I can say that SB is done pretty similarly to that, in a square+pair fashion, if not completely edge+pair+pair. SB doesn't really translate to ZZF2L because it uses a lot of Rw type moves, which, obviously, aren't applicable to ZZ. I think your best bet would be to try to learn all the tricks in there (there are some really cool ones), take advantage of any blocks and pairs, and see if you can find example sovles.

 

[AlphaSheep]

What I need more is tips on creating 3x2x1 blocks in the least moves possible, as my ZZF2L ranges from 20-30 moves, and I want to get that down to <20 moves. Anyone have any suggestions?

30 moves is a little excessive (assuming you're excluding EOLine), but 25 moves is about average, and 20 moves in a speedsolving situation is insanely efficient.

Post some examples of cases you really struggle to solve efficiently and I'm sure you'll get lots of advice on what you could have done differently.

Some general advice:

• Square then pair is most efficient around 95% of the time.
• Have a long look through the ZZ example solves thread and see what tricks you can pick up there.
• Do lots of linear FMC solves - that's where you count each move, take your time to plan each move, and aren't allowed to backtrack.
• D moves are still allowed. Have a look at keyhole. Inserts like DRUR'D' can be used to insert a corner in DLF or and edge in FR without affecting the other pieces. There are lots of variations on that.
• For the first block, take advantage of the open space on the other side - you don't have to waste moves restoring pieces there
• Not all F2L algs with F moves are bad. A really good example is F'RUR'U'R'FR. Go through CFOP F2L algs for cases you find inefficient and see if any short algs preserve EO.
• For really bad cases, try HARCS and see what solutions it gives you. Often you can use a non-obvious trick that makes you more efficient. Try to see the concept rather than the moves so that you can mirror it or adapt it to a different case on the fly.
• Practice all of this over and over and over.

 

[PapaSmurf]

A good trick is open slotting. When you solve a square, it’s actually more efficient to not do the final move and completely solve the square.

Example: U R U’ R2 U’ R’ U R’ U’ R’ U’ R U R2 U’ R2 U R’ U R U R’ U2 R U2

R U R’ U2 R U’ R2// square. Notice the R2 instead of an R at the end. This sets up the other pair nicely.
U R U2 R’ U’ R// last pair. Open slotting saved a move, as the solution for the second pair without openslotting is U2 R’ U’ R U’ R’ U R, but sometimes it can be more.

 

[Pyjam]

You can also play those games:
https://www.speedsolving.com/forum/threads/the-zz-example-solve-game.48190/
https://www.speedsolving.com/forum/threads/the-3x3x3-example-solve-thread.14345/
...where you'll find many examples. The Leor method uses blockbuilng too.

Also, the advices par AlphaSheep are very good.

 

[ZZeroingZZemdegs]

30 moves is a little excessive (assuming you're excluding EOLine), but 25 moves is about average, and 20 moves in a speedsolving situation is insanely efficient.

According to the SpeedSolving Wiki, though, the average move count for ZZF2L should be 22 moves, as it states that the average movecount when using ZBLL is 44. EOLine is 6 moves and ZBLL is 16 moves, which leaves a total of 22 moves for F2L.

 

[AlphaSheep]

According to the SpeedSolving Wiki, though, the average move count for ZZF2L should be 22 moves, as it states that the average movecount when using ZBLL is 44. EOLine is 6 moves and ZBLL is 16 moves, which leaves a total of 22 moves for F2L.

I think the numbers quoted in the wiki are not completely accurate.

The best estimates for I know of, based on strategy are by Lars Vandenbergh:

• Perfectly optimal both blocks simultaneously: 15.64 moves [source]. This is not even remotely humanly possible.
• Optimal blocks seperately: 19.72 of you have a fixed order, or 19.12 if you always choose the shortest block first [source]. A human could conceivably do this in a handful of special situations, but most of the time, it's not feasible in a speedsolve.
• Square then pair on one side, and then a square the pair on the side: 25.61 moves with a completly fixed order, 24.10 moves if you always do the same side first, but have no bias to the front or back, or 23.51 moves if you always pick the side with the shortest square [source]. This is definitely possible for a human to acheive.
• Finish cross, then solve 4 pairs (CFOP style) using <R, U, L, D>: 27.72 moves or using <R, U, L>: 29.18 moves based on 100 sample solves [source].

There's some other interesting reading here and a brief analysis using HARCS here.