[Help Thread] 4x4 Discussion

[genwin]

i used to do 1 at a time then i watched frank morris video guide for 2 at a time from bigcubes.com(its downloadable)... i watched it at least 20 times pausiing in between... its very very useful..

 

[cookingfat]

thanks everyone, I think I've got the hang of it now, I'll check out the frank morris vid too.

btw, to get free cubes, click the link in my sig. I got enough credit for a FREE 4x4x4 or 5x5x5 eastsheen in only 2 hours, no joke !!

 

[somerandomkidmike]

4x4 solving question

How many moves would you have to use for direct solving methods (centers last) or anything like the cage method to be as good as reduction?

How many moves do you all average for reduction or cage (or whatever method you use?

 

[rachmaninovian]

hehe, for 4x4 its around 125-150 depending on centres. 5x5 220-250 again, depending on centres. but since centres can be switched quite fast with niklas it doesnt really matter.

 

[somerandomkidmike]

how many moves do people use for reduction? for 4x4?

I'd like everybody to post how many turns they use for their method, and tell what method they use if possible (thanks in advance)

 

[ShadenSmith]

http://www.speedsolving.com/forum/showthread.php?t=3177&highlight=reduction+move+count


Try searching.

 

[somerandomkidmike]

I've seen this thread, but I wanted to know... how many moves would a direct solving method need to be "as good" as reduction?

 

[fanwuq]

I'd say just use reduction for 4x4.
Advantages:
(Less moves, if not about the same)
Better look ahead
More fingertrick friendly.

I get around 110-140 using Reduction and I suck at 4x4. I just make 1 pair at a time, LOL. I'm pretty sure better people can average around 120 moves per solve. I see nothing better about direction solving methods. They require slice moves and solving specific pieces, so it's not so easy to be color neutral.

 

[somerandomkidmike]

I'd say just use reduction for 4x4.
Advantages:
(Less moves, if not about the same)
Better look ahead
More fingertrick friendly.

I get around 110-140 using Reduction and I suck at 4x4. I just make 1 pair at a time, LOL. I'm pretty sure better people can average around 120 moves per solve. I see nothing better about direction solving methods. They require slice moves and solving specific pieces, so it's not so easy to be color neutral.

I use Rachmaninovian's method for the 4x4, and I use less moves than him. I average about 115 moves, and the look ahead is great. On a slow turning solve I average about 1minute 50 seconds, and on a regular turning solve I average about 1 minute 45 seconds.

I average the same number of turns as reduction as a lot of people (less in many cases)

Also, as for fingertricks I can do the center switching algorithm in the same amount of time as niklas (I just did it in 1.88 seconds, and I did 1.86 for niklas)

I still want to know what people think about this... if I could possibly bring down the average number of turns for a direct solving method down to 110 or less, do you think it could have as much potential as reduction?

 

[rachmaninovian]

eew, ur move count is amazing, i dunno how you do it lol. the center switching alg from jelinek's pages suck...niklas is better

EDIT: thinking about it, perhaps the only way you reduced the move count (that I can think of) is by pairing up like 2 or 3 more dedges before inserting? I used to do that until I decided that it is not the best thing to do; it is slightly slower and an average of 6-7 moves per dedge for the first 8 isn't too bad. say you can do one in 2.5s you can finish the first 8 dedges in 20s? I have done it before.

please clarify? haha maybe you have centre magic it is the slowest part of the solve, and needs A LOT of turns. 7-8 moves niklas with set up move for a single swap sucks really. obviously i switch 2 at the same time if they are in blocks.

Also, perhaps provide a break-up of your solves?
for me it WAS (i no longer practise 4x4)
first 2 centres - 5s
corners- 12s (recognition )
first 8 edges - 25s
finish cage - 12s
centres - 25s
total - 79s plus/minus 5s depending on cases.

 

[TMOY]

For me it's something like:
corners: 10 s
first 2 centers: 15 s (perhaps I should switch those two steps ? Not certain yet)
first 8 edges: 45 s (I suck at edges . I solve them 1 at a time except the last two, I haven't experimented 2 or 3 at a time yet)
finish cage: 15 s
last 4 centers: 20-25 s (There are definitely much faster ways to solve them than lots of noklaas )
Which adds up to 1:45-1:50.

 

[rachmaninovian]

hey but i solve one dedge at a time and TMOY i would want your centres mine is too inconsistent

 

[somerandomkidmike]

eew, ur move count is amazing, i dunno how you do it lol. the center switching alg from jelinek's pages suck...niklas is better

EDIT: thinking about it, perhaps the only way you reduced the move count (that I can think of) is by pairing up like 2 or 3 more dedges before inserting? I used to do that until I decided that it is not the best thing to do; it is slightly slower and an average of 6-7 moves per dedge for the first 8 isn't too bad. say you can do one in 2.5s you can finish the first 8 dedges in 20s? I have done it before.

please clarify? haha maybe you have centre magic it is the slowest part of the solve, and needs A LOT of turns. 7-8 moves niklas with set up move for a single swap sucks really. obviously i switch 2 at the same time if they are in blocks.

Also, perhaps provide a break-up of your solves?
for me it WAS (i no longer practise 4x4)
first 2 centres - 5s
corners- 12s (recognition )
first 8 edges - 25s
finish cage - 12s
centres - 25s
total - 79s plus/minus 5s depending on cases.

Well my solves I don't do one center at a time. I often do 2, and in some cases I have done as many as 8 in one. Also I pair up edges before doing the first layer. In fact, I blockbuild my first layer completely.

If there are 2 centers in a pair, I will solve them all at once. Maybe I just got a few lucky solves in a row, but I do try pairing using block building, and pair up a lot before solving if they are quick to solve cases

Here is my order of solving the 4x4

1) solve 2 opposite centers
2) Blockbuild the first layer minus one (paired) edge.
3) Solve the other corners in one algorithm (still learning the cases, but they are an average of 8 turns using the CF algorithms)
4) Solve the Redges minus 1 edge pair, and pair up the other two edge pairs
5) Solve the last 2 redges
6) Solve the midges
7) Centers

Also whenever I see an edge that can be paired up before finishing the corners, I will do it.

 

[Escher]

wait........

8 centres in one?
im probably not following your meaning, but arent there only 6 on a cuboid?

 

[rachmaninovian]

i think he meant cubies. doing M2 b2 M2 b2 does mess up 8 centre cubies =P

 

[somerandomkidmike]

Ok, so I just did an average of how many moves it took, and I averaged 120 turns using Rachmaninovian's method.

You can switch 8 centers by (Rr)2 E2 (Rr)2 E2

(I'm not sure about the notation for 4x4, but I think this may be right)

I always solve centers in opposites, because of these fast cases.


Anyway, if i knew all the cases I need to, then I would average less turns, so I will continue learning them (I'm making them, because I am a CF speedcuber).

 

[rachmaninovian]

what other cases? do share with me =P

anyway, although you will average less turns by doing certain things instead, speed might be compromised. I was advised by Kenneth that trying to solve as many pieces at a time to save moves WILL destroy lookahead.

yes solving centres in opposite is good...I do it too...but I just solve everything that comes in the way. hence direct solving =P but if I see that I can pair up a pair to set up a 1x1x2 block to switch to the opposite face, I will do it. you must remember that in a speedsolve, you will lose move counts by doing whatever that is fastest at the moment. I had a few sub 1:20 vids, and amazing the move counts were in the 150s, and when I do slow solves, I average 135-145 moves.

btw, solving adjacent centres for 5x5 is definately better. I can average almost 40s now for the last centres =P

and anyway, if you can do one niklas in like 3s with AUF, why do you care to optimize the move count for centres if everything can be done very fast anyway? per told me that he had a dislike for opposite centres simply because they are a tad slower to execute. haha if I remembered, yes, my centres averaged 23-25s. i really forgot but i don't remember going over 30s.

all in all, it is how fast you can do each step, not how low the move count can go for each step. anyway, for solving the last 2 edges of the first 2 faces, i think pairing up and solving both at the same time is slightly slower on average though faster on singles then my method of putting one at a time. consistency is also very important anyway =P

 

[TMOY]

On 4^3, centers can be done in two looks. First, put every cubie on the right pair of opposite centers. Second, solve both pairs of opposite centers at once.
That's 69 algorithms to learn (42 for the first step and 27 for the second), not counting setups. I'm still far from mastering all the cases. But it's definitely possible.

 

[somerandomkidmike]

I solve opposite centers for the 5x5 when it is faster (because of cases like (Rr)2 U2 (Rr)2 U2 (Rr2) (Ll)2 U2 (Ll)2 U2 (L2) and (Rr)2 U2 (Rr)2 U2 R2 Ff2 U2 (Rr)2 U2 f2 r2

I'll share the cases later, but I have to go.

 

[rachmaninovian]

anyway, does small letter in your notation mean double slice turns?