Why you should learn COLL for 4x4x4

[deco122392]

You're definitely right about the 2.5 bit... that's convinced me that it might be worthwhile to learn. Do you have a list of good COLL algs (for 4x4) somewhere?

I use the same ones as i do for 3x3 they all work enough and i have no problems comfort-wise with them.

http://jmbaum.110mb.com/coll.htm

of course you can modify the algs slightly for you personal style which shouldnt take more then a min of thought

[Lt-UnReaL]

After COLL, the probability of getting a PLL skip is 1/24, as opposed to 1/288 if you just did your usual alg. So you're 12 times as likely to get an LL skip with COLL as you are without, that's 1200% (not 120%, it's 1200%).

You mean PLL skip instead of LL skip?

I use about 14-15 COLLs right now for 3x3 and higher. I don't use them all mainly because of how bad the algorithms are compared to the best algorithm for that OLL. For example, let's say for a certain COLL case:

Sune - ~0.8 sec
G perm - ~2.0 sec

COLL for that Sune - ~2.1 sec
U perm - ~1.4 sec

Also, according to your data, COLL would be useful especially on any even numbered cube, 4x4 and higher.

[JLarsen]

Well as a Petrus user I find the oll's with all edges oriented incredibly nice, and I don't use COLL, but then again I've never really tried it. I've heard recog is actually slower and that makes sense to me too. Oh and on top of that the post above me says the algs are bad =\.

[fanwuq]

Well as a Petrus user I find the oll's with all edges oriented incredibly nice, and I don't use COLL, but then again I've never really tried it. I've heard recog is actually slower and that makes sense to me too. Oh and on top of that the post above me says the algs are bad =\.

I know all the Pi, Double Sune, and Triple Sune cases. They are all fast algs.
I know most of Headlights and Chameleon and they aren't bad either. All the COLL algs I know so far can be executed in less than 1.5 seconds. Many of them are simple 8 or 9 move commutators. I know the COLL algs for Sune, but I can't recognize them (Niklas, Sune, Bob, Buffy, and Anti-Buffy are pretty fast LUR algs, but I don't really like Doug Li's move.)

http://lar5.com/cube/yB.html

Edit:

Show me a case that "sucks hard." The average COLL should be faster than the average PLL. No COLL case is anywhere near as slow as N or E perms.

[DavidWoner]

1/12 of the time you have to do "the other PLL parity alg" (the one that swaps 2 adjacent edges) to get a U perm, otherwise you get an annoying Z or H perm, and fortunately enough for you, there are only 2 such cases and they're easy to recognize - opposite colors are opposite each other (try your best to understand this, it's really easy to recognize but not easy to describe )

Unless you have severe brain damage you should never get H-perm after PLL parity.

And I agree that a lot of COLL cases suck hard, but some are quite useful.

[Lord Voldemort]

Some people use an adjacent dedge swap.

[masterofthebass]

After COLL, the probability of getting a PLL skip is 1/24, as opposed to 1/288 if you just did your usual alg. So you're 12 times as likely to get an LL skip with COLL as you are without, that's 1200% (not 120%, it's 1200%).

You mean PLL skip instead of LL skip?

I use about 14-15 COLLs right now for 3x3 and higher. I don't use them all mainly because of how bad the algorithms are compared to the best algorithm for that OLL. For example, let's say for a certain COLL case:

Sune - ~0.8 sec
G perm - ~2.0 sec

COLL for that Sune - ~2.1 sec
U perm - ~1.4 sec

Also, according to your data, COLL would be useful especially on any even numbered cube, 4x4 and higher.

dude... sune COLL cases are the worst thing ever. Period.

[pjk]

Since learning COLL, I've given up on using it on 3x3.

Did you give up on Roux, Dan?

Using COLL on 4x4 is quite nice. There is really no point not to other than the fact that some of the cases are so crappy that it may be faster to not use COLL.

[masterofthebass]

I didn't give up on Roux. I learned all the CMLLs, and then filled in the gaps missing of COLL. Surprisingly quite a few algs carry over.

[blah]

After COLL, the probability of getting a PLL skip is 1/24, as opposed to 1/288 if you just did your usual alg. So you're 12 times as likely to get an LL skip with COLL as you are without, that's 1200% (not 120%, it's 1200%).

You mean PLL skip instead of LL skip?

Uh, yeah. Thanks for pointing that out As I've mentioned, it was 3 in the morning when I did this

1/12 of the time you have to do "the other PLL parity alg" (the one that swaps 2 adjacent edges) to get a U perm, otherwise you get an annoying Z or H perm, and fortunately enough for you, there are only 2 such cases and they're easy to recognize - opposite colors are opposite each other (try your best to understand this, it's really easy to recognize but not easy to describe )

Unless you have severe brain damage you should never get H-perm after PLL parity.

Yes, as a matter of fact, I did have severe brain damage due to severe lack of sleep last night. My bad. Thanks for pointing that out too

dude... sune COLL cases are the worst thing ever. Period.

To execute? I don't know, yet. To recognize? Not at all. It's definitely as easy as any other COLL case, well at least it is when I use my recognition system.

By the way I use RUL algs (mostly) for COLL, because 1. I learned COLL for OH and it's been tremendously useful, and 2. I prefer RUL algs to RFU ones on big cubes, I just lock up less.