cage/cage variants sub minute average potential?

[cubacca1972]

wells I took an average of 20 for the number of algs needed...right now it's at 4.25 algs. not too bad for short and fast algs eh? i'd probably learn a few more useful ones and stuff like that..
i'd do an average of 400 to make sure...happy waiting...xD

An average of 4.5 is very good. What is the worst case scenario possible with your center control system? Is it capped at less than 8?

Using longer algs will only make sense if they get lots done. If the net result is similar move counts to the short/fast set, but harder to execute, then short/fast will be the way to go.

BTW, I was trying to find your reference to the multi Niklas alg. What does it do?

For my system, now that I have fully recognized that I have the option of disregarding what happens on the adjacent faces (steps 1 and 2), I have found some fun stuff:

r' U2 r E2 r' U2 r E2

Out of all the permutations for solving opposite faces in step 3 and 4, 47% will be one of the easy cases (swap 1 center cubie, swap 2 adjacent on each face, swap whole centers).

Toss in the odd case where the first center gets solved in step one, and the total number of steps drops to 3 (or less).

 

[rachmaninovian]

the worst i had for yesterday was 6.

and by 1 alg i mean something that I can execute in one shot, like perhaps a combination and cancellations of 2 niklas? for example,
r U l' U' r' U l U' l' U r U' l U r' U
which cancels to
r U l' U' r' U2 r U' l U r' U

and again another disclaimer: I do not control centres all the time; it is not always feasable. I would say that centre control gives me more of better times rather than help me get out of trouble. It does help reduce bad cases sometimes though.

so basically the number of algs I know would automatically reduce the maximum number of algs needed to solve the centres, assuming that you do clever tricks.

 

[TMOY]

Note that most longer algs are nothing else than combinations of short algs, like your double niklaas for example. Same move count as the short algs (or lower with cancellations), exactly as easy to execute, and it saves some looks.

 

[rachmaninovian]

yea thats what i meant. the best thing abt niklas is only ONE setup move is needed always...so it is very very fast.

but my main purpose of centre control is to speed up the final centres so that I get easier cases that require less time recognizing and stuff like that =P

 

[cubacca1972]

Note that most longer algs are nothing else than combinations of short algs, like your double niklaas for example. Same move count as the short algs (or lower with cancellations), exactly as easy to execute, and it saves some looks.

My hope would be that with a search program, more direct solutions could be found which might be shorter.

Welding two shorter algorithms together certainly works, but can't be the only way to get the job done. A computer will simply be able to find them more effectively than we can.

Until such time, I am welding.

and by 1 alg i mean something that I can execute in one shot, like perhaps a combination and cancellations of 2 niklas? for example,
r U l' U' r' U l U' l' U r U' l U r' U
which cancels to
r U l' U' r' U2 r U' l U r' U

Thanks for the Niklas explanation. I was a bit slow on the uptake, and didn't recognize it as a Niklas, given the use of the inner layers.

 

[rachmaninovian]

ah wells my average seem to be improving slowly...though I get sub 1 solves several times a day...most of them being non lucky of course..

I still get stupid times because of stupid centres...after getting a 50.34s solve i got a 1:23.xx -.-

although I would say I do freestyle centres, I find myself solving one centre first before solving the rest half the time..lol..

cubacca1972: perhaps you should try ksolve =P if you know how to use it that is...i can't bother actually =P i'm busy generating more edges algs hahahaha..

anyway i have an average of 60 now...its at abt 4.08...so far 2 solves had 6 algs centres zzz...

 

[cubacca1972]

cubacca1972: perhaps you should try ksolve =P if you know how to use it that is...i can't bother actually =P i'm busy generating more edges algs hahahaha..

I would try ksolve, except for 2 things.

1. I can't seem to find a working link to it.

2. I am currently using someone elses computer, and don't have admin permission to install anything.

Once both issues are resolved I plan to check it out.

I am still not satisfied with the all 4 center cubies in an adjacent face swap. It just seems like there must be a smoother way.

 

[rachmaninovian]

u M' U2 M u' d' M' U2 M d - isnt this great? if you didn't realize...=P

 

[cubacca1972]

u M' U2 M u' d' M' U2 M d - isnt this great? if you didn't realize...=P

Yes, this one is very good. I do it as

r' E R2 E' r l E R2 E' l'

I have a slight aversion to algorithms with anti slice-like moves such as r l. I should just get over it with this algorithm, given how slick it is.

The last algorithm I am looking for in the 0 cubies solved/sorted set is 2 adjacent cubies on one adjacent face, 2 diagonal cubies on the other adjacent face.

 

[somerandomkidmike]

all of these algorithms are gonna help me sooooooooo much.... well when I get a new 4x4

 

[rachmaninovian]

u M' U2 M u' d' M' U2 M d - isnt this great? if you didn't realize...=P

Yes, this one is very good. I do it as

r' E R2 E' r l E R2 E' l'

I have a slight aversion to algorithms with anti slice-like moves such as r l. I should just get over it with this algorithm, given how slick it is.

The last algorithm I am looking for in the 0 cubies solved/sorted set is 2 adjacent cubies on one adjacent face, 2 diagonal cubies on the other adjacent face.

I wanted to do it that way...but the original one is faster using fingertricks. There is also little regrip for doing it the original way =)

 

[cubacca1972]

u M' U2 M u' d' M' U2 M d - isnt this great? if you didn't realize...=P

Yes, this one is very good. I do it as

r' E R2 E' r l E R2 E' l'

I have a slight aversion to algorithms with anti slice-like moves such as r l. I should just get over it with this algorithm, given how slick it is.

The last algorithm I am looking for in the 0 cubies solved/sorted set is 2 adjacent cubies on one adjacent face, 2 diagonal cubies on the other adjacent face.

I wanted to do it that way...but the original one is faster using fingertricks. There is also little regrip for doing it the original way =)

Gotta go where the speed is. IF the original works best for you, then stick with it. I am trying to keep cube rotations to a minimum. I already am busy unlearning my original adjacent face center algorithms so that I can adopt your Niklas style 3 cycles. That's why I adjusted the algorithm to suit the standard orientation. Although I am not that fast, I seem to be able to do my adjusted algorithm faster than the original.

BTW, what kind of 4x4x4 do you use? I use a Rubik's brand. Some of the layers turn smoothly, while some seem a bit stiffer. I also have an Eastsheen, but don't like it as much, as it feels too light and flimsy.

 

[rachmaninovian]

mefferts ftw =) lube with silicone oil for best effects, IMHO.

for faster turning speed, perhaps practise the algs that you use for everything? i do time attacks for the 16 cases of the last 2 dedges =) it is always important to do all the dead parts fast because you have to waste time thinking and processing during the intuitive parts..

EDIT: 1:01.56 average, 10 of 12, (and with 4 sub 1s xD) but with so many smooth solves i don't think its a good representative of my real average which i estimate to be around 1:05 +/- 2s...i would call this average lucky..

 

[masterofthebass]

Average: 57.10
Standard Deviation: 3.02
Best Time: 52.40
Worst Time: 1:02.45
Individual Times: 53.02, 52.71, 1:01.37, 59.62, 57.49, 1:00.84, 59.61, (52.40), 56.65, 54.90, (1:02.45), 54.81

(k4)

 

[cubacca1972]

mefferts ftw =) lube with silicone oil for best effects, IMHO.

for faster turning speed, perhaps practise the algs that you use for everything? i do time attacks for the 16 cases of the last 2 dedges =) it is always important to do all the dead parts fast because you have to waste time thinking and processing during the intuitive parts..

EDIT: 1:01.56 average, 10 of 12, (and with 4 sub 1s xD) but with so many smooth solves i don't think its a good representative of my real average which i estimate to be around 1:05 +/- 2s...i would call this average lucky..

Ahh, Meffert's. I have bought a bunch of stuff from there: Skewb, Skewb Ultimate, Pyraminx, Eastsheen 5x5x5, 4x4x4, 2x2x2, Megaminx, Dogic. In retrospect, I enjoyed acquiring them more than having them. At least some of them. I think I would trade a few of those for a Meffert's 4x4x4. Oh well.

I think I will keep looking for algorithms for a while before working on speed.

I did have a thought in terms of the 4 step centers. Disregarding center cubies that are already solved/sorted once the cage is completed, there are 20 different ways that 1 to 8 center cubies (the ones that need to be sorted) can appear on the adjacent faces. This leads to all sorts of interesting lines of study, assuming that I ever get a complete set of algorithms. If you have several options to sort/solve in step 1, which do you pick? What will be left over for step 2? Which center should you start with? Would it ever be worth it to disregard all (or some) sorted cubies in step 1, because it would lead to a more ideal step 2?

My brain hurts.

 

[rachmaninovian]

Average: 57.10
Standard Deviation: 3.02
Best Time: 52.40
Worst Time: 1:02.45
Individual Times: 53.02, 52.71, 1:01.37, 59.62, 57.49, 1:00.84, 59.61, (52.40), 56.65, 54.90, (1:02.45), 54.81

(k4)

=(

EDIT: yay 59.74 average. i would say a very lucky average...for one of the sub 1s all I did for the middle edges was r2 U2 M2 U2..
rolling average of course...after this average as usual my concentration went down...and a 1:08 average returned..

 

[somerandomkidmike]

Average: 57.10
Standard Deviation: 3.02
Best Time: 52.40
Worst Time: 1:02.45
Individual Times: 53.02, 52.71, 1:01.37, 59.62, 57.49, 1:00.84, 59.61, (52.40), 56.65, 54.90, (1:02.45), 54.81

(k4)

=(

EDIT: yay 59.74 average. i would say a very lucky average...for one of the sub 1s all I did for the middle edges was r2 U2 M2 U2..
rolling average of course...after this average as usual my concentration went down...and a 1:08 average returned..

Congrats! Now we'll have to see if sub-50 is possible for an average!

 

[cubacca1972]

EDIT: yay 59.74 average. i would say a very lucky average...for one of the sub 1s all I did for the middle edges was r2 U2 M2 U2..
rolling average of course...after this average as usual my concentration went down...and a 1:08 average returned..

I don't subscribe to the luck thing. If The event is so improbable, it won't come up often enough in an average of 10 to matter that much. I mean, you don't get to shrug off really "unlucky" solves in the context of taking an average either, so I think luck will tend to cancel out. You still have to kill on all the luck-neutral solves in the average.

Nice work so far.

 

[rachmaninovian]

explain the 1:08 average for the next 12 xD and no sub 1s too =P
wells it is difficult to sub 1 properly...lol..need lots of concentration..
and luck? I had a lot of 3 alg finishes, and for the next average i had a lot of 5 alg finishes...and the edges were much harder in the second average..

 

[cubacca1972]

explain the 1:08 average for the next 12 xD and no sub 1s too =P
wells it is difficult to sub 1 properly...lol..need lots of concentration..
and luck? I had a lot of 3 alg finishes, and for the next average i had a lot of 5 alg finishes...and the edges were much harder in the second average..

Best guess, as I can't view your whole method without crashing my browser:

With edges, I am assuming you mean the l and r edges after completing 2 opposite faces. Depending on how many algorithms you use, and how many cases you need to know to achieve center control, I wonder about the amount of stopping/slowing down needed to decide what to do next.

With centers, I am guessing that that you ended up with lots of 5 step finishes along with lots of misses while attempting center control during edge completion.

Disregarding center control, how do you finish your edges? I am guessing that you solve one edge pair, then an adjacent edge pair, then the last two edge pairs in one algorithm. Is this correct?

As an experiment, try finishing your cage with your fastest algorithms only.
See if this cuts down any potential time gaps between algorithms (fewer choices to make, less brainpower needed to decide what to do next). Then solve your centers as you usually do. See if this leads to any siginificant increase in number of steps or degree of ugliness in cases.