ZZ-CT vs MGLS-Z

[Hammer]

Greetings,

I've been cubing for 3 years and average 20 seconds with CFOP, with a pb of 12.93. I recently switched to ZZ, and I'm wondering which variant to use. I watched Chris Tran's seminar at Nationals and immediately started learning TSLE. However, I realized that you always need to leave out the FR slot, which kinda sucks. Then I realized that you could do it from multiple angles, but that seems like a lot of work. Then I tried the beginners guide (22 algs) which you can find here:

https://www.speedsolving.com/forum/threads/zz-ct-for-beginners-22-algs-and-zz-mgct.62066/

And while I was learning the algs for this, I came across MGLS-Z, which seems a bit better, since it forces PLL instead of TTLL, which means a higher skip rate. However, the recognition isn't great. If I want to get a sub-10 single, which method gives me a better chance? I should add that my TPS isn't that great, I can't even sub-1 a U perm.

 

[xyzzy]

Focusing on skip rates seems to be completely misguided. There are two parts to getting fast singles: being faster on average, and increasing your variance. The second part is kind of iffy and there isn't really much useful advice that can be given, because it's more likely to backfire than to help.

Which leaves the first part: getting faster overall. My own tps is pretty mediocre (I can't even sub-2 any EPLL consistently)… and yet I'm somehow sub-19. At your current times, you don't need moar fasts to get faster; what you need is to reduce pauses. (Reducing pauses is something I also have to work on, of course.)

Also, fancy LSLL systems won't help if the bottleneck is in your F2L; if you can't consistently sub-10 your F2L, getting a sub-10 single for a full solve is out of the question.

Edit: Also also, CLS has a roughly comparable alg count to TSLE and if you want to be LS-neutral, you'd have to learn four times as many cases, just as with TSLE. If your original reason for not wanting to do TSLE is that it has way too many cases, CLS is even worse.

 

[Hammer]

Focusing on skip rates seems to be completely misguided. There are two parts to getting fast singles: being faster on average, and increasing your variance. The second part is kind of iffy and there isn't really much useful advice that can be given, because it's more likely to backfire than to help.

Which leaves the first part: getting faster overall. My own tps is pretty mediocre (I can't even sub-2 any EPLL consistently)… and yet I'm somehow sub-19. At your current times, you don't need moar fasts to get faster; what you need is to reduce pauses. (Reducing pauses is something I also have to work on, of course.)

Also, fancy LSLL systems won't help if the bottleneck is in your F2L; if you can't consistently sub-10 your F2L, getting a sub-10 single for a full solve is out of the question.

Edit: Also also, CLS has a roughly comparable alg count to TSLE and if you want to be LS-neutral, you'd have to learn four times as many cases, just as with TSLE. If your original reason for not wanting to do TSLE is that it has way too many cases, CLS is even worse.

Thanks for your insightful reply! I am now going to focus mainly on sub-10ing my f2l.

Lastly, if I don't care about alg count (i love memorizing algorithms) and all i care about is recognition speed and move count, which LS+LL system should I use? Which one do you use and what is your pb for zz method? And also, do you recommend FreeFop?

And also, i dont think for CLS you need to learn 4 times the algorithms, because your last f2l edge is already solved, so if i do a y axis rotation to bring the last slot to fr, it won't affect anything.

 

[xyzzy]

Lastly, if I don't care about alg count (i love memorizing algorithms) and all i care about is recognition speed and move count, which LS+LL system should I use? Which one do you use and what is your pb for zz method? And also, do you recommend FreeFop?

For the last slot, I use a few intuitive VHLS cases and 4 or so ZBLS cases. For the last layer, around 50% of it is OLL/PLL, around 40% is COLL/EPLL, and the rest is ZBLL or doing weird stuff. My PB with ZZ is 11.46 on a sim, 13.90 on a physical cube.

I don't recommend FreeFOP because blockbuilding on the fly isn't easy, but I find the freestyle approach a bit less boring than just building one pair after another. I probably could be faster if I switched to doing CFOP properly, but that wouldn't be fun.

And also, i dont think for CLS you need to learn 4 times the algorithms, because your last f2l edge is already solved, so if i do a y axis rotation to bring the last slot to fr, it won't affect anything.

Oh right, I forgot about that. You can also ADF instead of rotating (which might or might not be faster).

 

[GenTheSnail]

Lastly, if I don't care about alg count (i love memorizing algorithms) and all i care about is recognition speed and move count, which LS+LL system should I use? Which one do you use and what is your pb for zz method?

ZZ-a (ZBLL) is considered the best in the long run, and if you like learning algorithms, it shouldn't be to much of a burden. ZZ-b (phasing so EPLL is always H/Z) is comparable to TSLE and TTLL.
There were some arguments for ZZ-b over ZZ-CT, and you might want to read those before you dive too deep into a LS-LL method.
An argument starts here, and continues to the end of the thread. It's dead too, btw, so please don't bump it.
mDiPalma makes another post about it here, and you might want to read that.

I use full COLL/EPLL and about 20ish ZBLLs aside from my COLL skips. I plan on learning full ZBLL however. 2H PB is 9.33, OH is 11.06, Feet is 38.21. All in my signature.

 

[genericcuber666]

I think zzll is best, easy to learn with realy good expandability

 

[Hammer]

For the last slot, I use a few intuitive VHLS cases and 4 or so ZBLS cases. For the last layer, around 50% of it is OLL/PLL, around 40% is COLL/EPLL, and the rest is ZBLL or doing weird stuff. My PB with ZZ is 11.46 on a sim, 13.90 on a physical cube.

I don't recommend FreeFOP because blockbuilding on the fly isn't easy, but I find the freestyle approach a bit less boring than just building one pair after another. I probably could be faster if I switched to doing CFOP properly, but that wouldn't be fun.



Oh right, I forgot about that. You can also ADF instead of rotating (which might or might not be faster).

Yes, but remember, I don't use CFOP, I use ZZ, so the edges are already pre-oriented.