[Help Thread] Square-1 Discussion

[uberCuber]

I use the ones listed here

 

[Lid]

I use the ones listed here

Never liked those.
Here is my list.

 

[uberCuber]

Never liked those.
Here is my list.

I don't like those at all, so many >3 moves. To each his own then

 

[bobthegiraffemonkey]

Think this is the best place for this? Not sure it fits exactly but last time I posted elsewhere and had it moved here.

Parity + cubeshape in inspection:

[video=youtube_share;ldBm-Bnwk9g]http://youtu.be/ldBm-Bnwk9g[/video]

Damn I'm a good cameraman.

This is an introduction to correcting parity during inspection. I think this is a viable method and I intend to learn it now and use it in comp (but only once I've learned all of it). Feel free to ask questions, I will post it on the Speedsolving forum so you can reply there too. If you take sq-1 seriously and actually want to use this, I'll happily help as much as I can, so feel free to PM me there if this applies to you.

This covers the basics, there's a few things that can complicate this a little if you're not careful. At the very least, learning some easy cases for this could be helpful, rather than learning the full system. I will be tracking my progress in my signature.

Also, forgot to say in the video but for square/square scrambles (lol), I'd just use 'normal' tracing since that's easy enough. Luckily it was easy to get my sq-1 orientation to match my BLD orientation.

 

[EMI]

I'm afraid I don't quite understand what you are doing Ok so I see you have a specific order of edges and a specific order of corners. And I understand that for the case you are mainly talking about you have to do a 0,2 to fix parity if necessary. Now, what exactly are you counting? (I'm just interested in this specific case, as it seems to be the easiest one at the beginning)

 

[blade740]

If I understand correctly, you're still counting cycles as in the BLD method, you just learned a "default" even permutation for this (and eventually, I assume, each) shape, so you don't have to bother tracing back to cubeshape or memorize translation matrices.

 

[bobthegiraffemonkey]

Now, what exactly are you counting?

I'm counting how many 'targets are needed' to solve the pieces. Or rather, whether it is even or odd.

Consider (scramble in your BLD orientation): R' B' U2 D F' U2 F2 U' B' L' D2 L B2 R' B2 L B2 D2 F2 R'

Now, looking at corners with ULB buffer and tracing relative to the natural solved positions of each piece we get:
RFU (1)
RUB (2)
LFD (1)
UFL (2) - new cycle
BLD (1) - lolBLD
DFR (2)
UFL (1)

So we get odd parity for the corners. Do this for the edges and corners on sq-1, tracing relative to the chosen reference positions of the pieces (as shown on my black sq-1) and you get the parity. In the example I ended on (2), which is even parity and means that for that case I solve 'normally' to get even parity once I get to cubeshape - which I did.

Does that help?

E:

If I understand correctly, you're still counting cycles as in the BLD method, you just learned a "default" even permutation for this (and eventually, I assume, each) shape, so you don't have to bother tracing back to cubeshape or memorize translation matrices.

Almost. I've not learned even permutation for each case, I'm using the same reference for a particular shape of a layer regardless of what the other layer is or how I solve that case. I instead keep track for each case whether the 'normal' way of solving maintains or switches parity relative to my reference scheme.

If I get even parity tracing for the group of 8 edges case I solve normally, if I get 'even parity' for scallop/scallop I switch parity then start going to cubeshape.

 

[bobthegiraffemonkey]

CO/EO/CP/EP isn't very fun IMO. Are there any cool, fun and not too difficult to learn methods for solving after cubeshape? I'm thinking something similar to corners first could be nice.

 

[blade740]

Roux is pretty easy - blockbuild two CEC blocks on the D layer, then permute U corners, then Last 6 Edges. If parity is done ahead of time then 1-look L6E might be feasible.

 

[Akash Rupela]

I just did an execution only average of 15. This doesnt include any recognition between steps as that is another thing but includes AUFs and E slice flip during EP, I average about 21
cubeshape 4.71
co 1.06
eo 2.15 (full EO)
cp 3.45 (i use full parity CP)
ep 5.48
sum 16.85 (i.e 4 sec of transitions).

Can somebody please tell me what to work on, and if there is anything that is good, comments on all parts would be appreciated.

 

[EMI]

I think cubeshape could be faster, I recommend just looking into some kind of alg table every now and then.
Also personally I think there should be little to no pause at all between cubeshape and corner seperation (try to 2 gen a lot, and then look ahead into the corner seperation case you are going to get). Maybe you are already doing so, I can't tell from these splits

 

[vcuber13]

I agree cubeshape could be faster. Also, if you don't already, I would suggest using edge separation algs that do not effect CP. This allows you to predict the CP while doing EO, which helps a lot with parity CP.

 

[Jimmy Liu]

I use the statistic from my last official competition:

cubeshape 3
co 1.2
eo 1.2~2.2 (because I skipped 2 EOs)
cp 2.7
ep 5.3~6
sum 13.4~15.1

Maybe predicting algs is important

 

[Akash Rupela]

I think cubeshape could be faster, I recommend just looking into some kind of alg table every now and then.
Also personally I think there should be little to no pause at all between cubeshape and corner seperation (try to 2 gen a lot, and then look ahead into the corner seperation case you are going to get). Maybe you are already doing so, I can't tell from these splits

Thanks for your comment. It is very motivating to see your turn speed. My lookahead is pretty non-existent at the moment. I will surely try to implement what you are saying in my solves. Also, currently i have been just using the advanced cubeshape video by David Woner. I will then surely start with an alg table. I have also downloaded Raul's full cubeshape video for reference . I hope i can improve quite a bit there once i am able to get all cases in muscle memory.

I agree cubeshape could be faster. Also, if you don't already, I would suggest using edge separation algs that do not effect CP. This allows you to predict the CP while doing EO, which helps a lot with parity CP.

Yeah. I was lucky to have learnt algs unconsciously that don't affect CP. However, the main time goes in recognising parity rather than CP.

I use the statistic from my last official competition:

cubeshape 3
co 1.2
eo 1.2~2.2 (because I skipped 2 EOs)
cp 2.7
ep 5.3~6
sum 13.4~15.1

Maybe predicting algs is important

Thankyou very much for the motivating comment. that cubeshape CO and EO is amazing. Do you use Parity during CP or EP?

 

[Jimmy Liu]

Thankyou very much for the motivating comment. that cubeshape CO and EO is amazing. Do you use Parity during CP or EP?

I don't use parity CP, while I only know 7 EPs which are adj/-, adj/adj, op/op, U+/adj, adj/U+, U-/adj, adj/U-.

 

[Akash Rupela]

I don't use parity CP, while I only know 7 EPs which are adj/-, adj/adj, op/op, U+/adj, adj/U+, U-/adj, adj/U-.

That explains your skewed splits. I am trying to imagine how good you would get if you learnt some more EPs

 

[the super cuber]

can anyone tell me an alg for opp-opp ep and adj-adj ep that flips e slice too?

thanks

 

[vcuber13]

1 / 5 -1 / -5 1 /
-2 / 3 / -1 -1 / 4 1 / 6 /

 

[rollerking321]

U perms and Z perm which flip the middle layer ?

 

[blade740]

For Z I use 1,0/-1,-1/3,0/1,1/-3,0/5,-1/6,0/.

It's basically just the standard M2 U M2 U' M2 alg, with a /6,0/6,0/ cancelled into the end.

For U perms ( in fact, for most algs ) i use similar cancellations.