Random Blindfold Cubing Discussion

[Fábio De'Rose]

You could aim for a 50% success rate and take your averages from the successes.

My success rate (across the last 100 solves) is around this figure, sometimes a bit higher. Right now I scored 53/100, out of which 18 are above 3min.

I'm actually trying to devise a way of measuring the mean of the last 100 successful solves on my CsTimer session, then rounding it up to have an estimate.

 

[Christopher Cabrera]

My success rate (across the last 100 solves) is around this figure, sometimes a bit higher. Right now I scored 53/100, out of which 18 are above 3min.

I'm actually trying to devise a way of measuring the mean of the last 100 successful solves on my CsTimer session, then rounding it up to have an estimate.

3 BLD is weird because it is scored by a single rather than an average. I personally use more of a "feel" metric rather than any numeric system. I just ask myself "If I had to do a mean of 3 right now what would likely be my fastest single?" Even though my pb is a 1:59. I feel like my average result would end up being closer to 2:20.

 

[lucarubik]

if you select your dnfs as dnfs cs timer calculates the average of the rest doesnt it
i should be missing something

 

[mark49152]

IMHO traditional averages like ao12 are not much use in BLD because they don't happen often and that level of accuracy usually requires compromising on speed.

What I do is delete the worst 50% of every session and take the average of the remainder. If my accuracy is 50% then I'm only deleting DNFs. If I'm having a good day and my accuracy is 75% then I'm rewarded by being able to delete the slowest 1/3rd of my successes as well, rather than penalised by having to count slow solves that I could have just DNFed. If my accuracy is below 50% then the session is a DNF.

Obviously it's not fair to compare these results against averages that others take using a different formula, but it's a metric that has worked consistently well for me in tracking my own progress.

 

[Fábio De'Rose]

IMHO traditional averages like ao12 are not much use in BLD because they don't happen often and that level of accuracy usually requires compromising on speed.

Well, ain't that accurate as can be. I actually got an Ao12 today, and rolled it for a few solves thereafter.

It was steadily on the low 3s and I definitely was playing much safer (and probably a bit tired already as well, lol. Did almost 100 BLD solves today, so there's that).

Won't necessarily claim sub-3 status yet, but I'm very close.

 

[FastCubeMaster]

LOL

DNF(1:34.80) R2 U2 B2 U F2 L2 D L2 U' L2 U' L R2 U2 F' R2 U B F' L B Fw Uw2

E: lol the scramble

 

[1973486]

Another "solve" with comms. This has parity but I immediately noticed the easy edges so did it anyway. You can switch the order of the second and third comms to include two U perms.

This probably isn't helping me to be able to use comms in actual solves but it's fun anyway...

EDIT: 7 comms+flip

 

[lucarubik]

Another "solve" with comms. This has parity but I immediately noticed the easy edges so did it anyway. You can switch the order of the second and third comms to include two U perms.

This probably isn't helping me to be able to use comms in actual solves but it's fun anyway...

another offtopic question, do you guys know how to use the simplify tool from that site?
i want to expand the comms in my list so i can get statistics for move counts, but i will keep it in a commutator way if i cant reverse it
I cant manage to make the program expand algs with wide turns/ rotations in the set up, no matter how i write them, any tip?
actually never mind, im having lots of problems not just two, ill use that tool when it works propperly, if it ever does

 

[pinser]

another offtopic question, do you guys know how to use the simplify tool from that site?
i want to expand the comms in my list so i can get statistics for move counts, but i will keep it in a commutator way if i cant reverse it
I cant manage to make the program expand algs with wide turns/ rotations in the set up, no matter how i write them, any tip?
actually never mind, im having lots of problems not just two, ill use that tool when it works propperly, if it ever does

Comm: [x R': [U2, R' D2 R]]
Expand: [x R': [U2, R' D2 R]] ---> x R' U2 R' D2 R U2' R' D2' R R x'
Simplify: x R' U2 R' D2 R U2' R' D2' R R x' ---> x R' U2 R' D2 R U2' R' D2' R2 x' (it converted the R R to R2)
I'm not sure what you mean by "make the program expand algs with wide turns/ rotations in the set up," can you give an example? If you mean make it rewrite x R' as Lw', I don't think it can do that. Other programs can, though

 

[Gold Cuber]

Since this is ment to be random, I CAN SOLVE A RRUBIKS CUBE BLD!

 

[lucarubik]

Comm: [x R': [U2, R' D2 R]]
Expand: [x R': [U2, R' D2 R]] ---> x R' U2 R' D2 R U2' R' D2' R R x'
Simplify: x R' U2 R' D2 R U2' R' D2' R R x' ---> x R' U2 R' D2 R U2' R' D2' R2 x' (it converted the R R to R2)
I'm not sure what you mean by "make the program expand algs with wide turns/ rotations in the set up," can you give an example? If you mean make it rewrite x R' as Lw', I don't think it can do that. Other programs can, though

Spoiler: cool f***ing story

idk why i thought if you wrote a comm in a line whatever you wrote before it the program was considering it the set up, i do know why actually; my first comms of the list i used to test it, UR buffer, dont have set ups or the set up of one would add up with the set up of the next one, either way... corners seemed to stay unscrambled
but then i noticed at the end of the list the cube was pretty much scrambled right and I saw after a wide turn set up commutator (wich are actually most in that list) it didnt undo the setup, and i thought the problem was the wide turn, while it was never considering it a set up in the first place, not for wide turns, not for single turns
wow that was a cool f***ing story

so i need to write the [ signes before and after the set up too, and the double dot, shouldnt be neither hard nor tedius to do with excell, the oppening [ actually...
anyway thanks, i was convinced the problem was the l/x/Lw turns for some reason

what i thought simplify did is to write it back in commutator form
is there a way too do that, with this tool or any other?
if there is not i rather keep them in commutator form too, you never know, also it would be cool to see the break down of some 2 gen algs, although i guess cancellations would make this thing waaaaay harder than expanding for a program

 

[pinser]

Spoiler: cool f***ing story

idk why i thought if you wrote a comm in a line whatever you wrote before it the program was considering it the set up, i do know why actually; my first comms of the list i used to test it, UR buffer, dont have set ups or the set up of one would add up with the set up of the next one, either way... corners seemed to stay unscrambled
but then i noticed at the end of the list the cube was pretty much scrambled right and I saw after a wide turn set up commutator (wich are actually most in that list) it didnt undo the setup, and i thought the problem was the wide turn, while it was never considering it a set up in the first place, not for wide turns, not for single turns
wow that was a cool f***ing story

so i need to write the [ signes before and after the set up too, and the double dot, shouldnt be neither hard nor tedius to do with excell, the oppening [ actually...
anyway thanks, i was convinced the problem was the l/x/Lw turns for some reason

what i thought simplify did is to write it back in commutator form
is there a way too do that, with this tool or any other?
if there is not i rather keep them in commutator form too, you never know, also it would be cool to see the break down of some 2 gen algs, although i guess cancellations would make this thing waaaaay harder than expanding for a program

Yes, a lot of people have trouble with proper commutator and conjugate notation. I can name a few I know *coughcough*
Conjugate: [A : B] = A B A' (setup moves are usually expressed as conjugates)
Commutator: [A, B] = A B A' B'
Putting them together, [A : [B, C]] = A B C B' C' A', which is what you will mostly be using for 3-style comms

I'm not aware of any software that will convert an alg to a commutator, but if you want a list of your comms in commutator and expanded form, there's always ctrl+a ctrl+c ctrl+v...

 

[lucarubik]

Yes, a lot of people have trouble with proper commutator and conjugate notation. I can name a few I know *coughcough*
Conjugate: [A : B] = A B A' (setup moves are usually expressed as conjugates)
Commutator: [A, B] = A B A' B'
Putting them together, [A : [B, C]] = A B C B' C' A', which is what you will mostly be using for 3-style comms

I'm not aware of any software that will convert an alg to a commutator, but if you want a list of your comms in commutator and expanded form, there's always ctrl+a ctrl+c ctrl+v...

oh i didnt have time to have trouble, i just never knew, anyway this was more tedious than hard but its done
UBL 9.93 12.42
UB 9.01 9.97
UBR 9.93 12.39
UR 9.41 10.49
i think i need to reconsider some 5 move insertions i use for corners R2 D R2 D' R2. U, move count doesnt mean anything but still
btw

Spoiler: really micorsoft

=LARGO(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR(SUSTITUIR([@algorithm];"[";"");"]";"");",";"");"'";"");" ";"");"x";"");"y";"");"z";"");"2";""))

im planning on doing a chart for corner twists too, using free buffer couple of comms algs or just algs like sune+ Uperm, maybe one for edges too

 

[lucarubik]

I've updated the lists grouping also edges, 2 edge flips and 2 and 3 corner twists cases by (isometry?)
feel free to use it if you can, to me this means i have way less cases to analyze with cubexplorer

 

[Daniel Lin]

ok i finally figured out how kaijun solves parity

he does DFR>last corner target>DBR
and DF>last edge target>DB

then x2 F perm

really inefficient, but kaijun has good tps....

also the x2 explains why half of his cubes in multi are flipped
@sigalig

 

[h2f]

ok i finally figured out how kaijun solves parity

he does DFR>last corner target>DBR
and DF>last edge target>DB

then x2 F perm

really inefficient, but kaijun has good tps....

also the x2 explains why half of his cubes in multi are flipped
@sigalig

That's intresting. It looks like his 3style is expanded version of M2/R2 method which I guess is rare. I know only one person who did M2/R2 and expanded it to 3style. But he used FD and FDR as buffers and changed FDR to UFR. Fperm as a parity used Grzegorz Jałocha - his buffers were UF and UBR.

 

[Fábio De'Rose]

CsTimer hates me:

Generated By csTimer on 2017-5-12
single: 1:53.27

Time List:
1. 1:53.26 [39.50] B L2 B2 U R2 F2 D' L2 U' B2 R' F U2 R' B' R F2 R2 D' Fw'

Why so many twisted corners? Like, seriously. My execution suffered heaps.

 

[kake123]

CsTimer hates me:

Generated By csTimer on 2017-5-12
single: 1:53.27

Time List:
1. 1:53.26 [39.50] B L2 B2 U R2 F2 D' L2 U' B2 R' F U2 R' B' R F2 R2 D' Fw'

Why so many twisted corners? Like, seriously. My execution suffered heaps.

Not too much of an issue for me though

L (U R U' R')*2 L' (R U R' U')*2 [Solves twisted corner at DBL]
R U' L' (U R' U') (R U R' U') L (U R U' R') (U) [Solves twisted corner at UFR]
(U') R U' L' (U R' U') (R U R' U') L (U R U' R') (U2) [Solves twisted corners at UFL and UBR, Note that this won't work if the corners are twisted in a different way]

 

[T1_M0]

Hi! I can't figure out how to fix midge parity for 5bld. My solving order is centers -> wings -> midges -> corners. I tried the normal 3x3 parity alg but that of course breaks the two edge blocks.

I'm also not quite sure if the 4x4 wing parity works on 5x5 without damaging the midges. I've yet never tried 5bld but with some tips I could be on my way.

 

[FastCubeMaster]

Can someone tell me the pros and cons of having DFR as your buffer using 3style? I know it involves more 'R' 'U' 'D' algs but anything else?