• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 35,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

F2L Edge Orientation guide

Joined
Jul 23, 2008
Messages
3,375
Likes
75
WCA
2008KINN01
YouTube
RowanKinneavy
Thread starter #1
A bit like the last thread I made, this guide is going to be text-based and the thinking behind each solve will be more complicated than that which anybody does in a normal speedsolve.

The basic idea of this post is to teach you how to recognise the orientation of f2l edges and also how to effect them so that you can reduce the number of rotations in your solves.
It's actually possible to reduce your f2l rotation average to less than one using this and with quite a few extra tricks (Breandan Vallance's old style is an extreme example).

Scramble WCA (green front, white top) and we're following cross on yellow (deal with it):

Example 1: L R' D R2 U L' B F2 R2 L D' U L' R' U2 R' D B R2 L2 D R2 U B2 U'
Cross: y F L' D L F D2

Here we recognise the orientation of the f2l edges:
First, know/check the colour of the centre pieces on L/R.
Second, look at each f2l edge - the important sticker on each is the one that matches one of the L/R centre pieces.

If this important sticker is either on the U layer orbit (that is, in any of FU, RU, BU, LU positions),
OR
on the L/R face (RU, RF, RB and the mirror),

then it's 'correct' - that is, you can solve the pair it belongs to with with just <L, U, R> moves.

If it's the inverse of these rules - on the U face or on the R orbit - then it's going to require a rotation, a fancy trick, or turns using the F or B faces.

So we apply this to the scramble so far. Since we already rotated before cross, the 'important stickers' for the f2l edges are the green/blue ones. We then check their orientation in relation to the rules we outlined earlier -
BL edge: correct since the sticker is on the L face
FL edge: " "
UL edge: " "
UR edge: incorrect, since the important sticker (here it's green) is on the U face.

And thus the UR edge is a problem we have to consider...

The interesting thing about EO is that we can now do a few different approaches:

Option 1 - flip the bad edge and gogo EOCross style:

U R' F R F' (orients the 'bad' edge and doesn't affect any other f2l edges)
U R' U2 R U' R' U R
R L' U2 R' U L (funky but relatively pointless trick)
U' L U L' U2 L U L'

Option 2 - solve the oriented pairs until you have to rotate (or until it's best to):

U2 R U' R' U R' U' R
L U L' U2 L U L'
y' - here we can rotate because the blue/red sticker in FL can also be solved in <R, U, L> post-rotation - if you don't know, work out why.
R U R2' U' R
U R U R'

Option 3 - abuse some more edge orientation/rotation rules:

Here we can add some more rules to our repertoire -
a) if a 'good' edge is in the U layer, a rotation will make it 'bad'.
b) if a 'bad' edge is in the U layer, a rotation will make it 'good'.
c) if either are in a slot, a rotation will not affect it's orientation.

Knowing this, we can rotate much earlier on:

U2 R U' R' U R' U' R (solving the only good edge on the U layer)
y' L' U L U2 L' U L (the other good edges are in slots, the bad edge in the U layer still)
R U' R' U R U' R'
U' R U2 R2' U' R2 U' R'

.

All of the above are viable, it really just depends on the scramble.

One thing to be aware of however - nobody properly using ZZ does EOCross - the average optimal f2l movecount is boosted from something like 27 to 36.
For this reason be careful of how you approach f2l EO - don't go crazy and effect it all the time - although getting fewer rotations is nice, beware of the fact you can seriously increase your movecount.


Another little interesting thing:
If a good edge is in a slot in it's correct layer, then it can be solved <R, U, L> regardless of rotations, since in either case it's 'important sticker' is correct relative to whichever centre you care about.
Using this knowledge, we can change our f2l very slightly to create these kind of situations more often than we would otherwise, and thus be free to use our rotations for more positive ends. *

Now I really have to go back to the work I was avoiding but I'll post plenty more examples later on today (I will update with a few more examples in my older f2l thread at the same time).

For anybody wanting to learn more about EO and EO detection visit Conrad Rider's site here. You can also learn a bit more about reducing the cube to move groups by reading parts of Mackys 3OP guide.
I strongly recommend Fridrich users to investigate the different steps of at least each main method (ZZ, Petrus, Roux) if not the more obscure ones (Heise, Waterman etc).
The more you understand blockbuilding, edge orientation, different LL methods etc etc the more you understand your own f2l.

Don't ever forget that Fridrich f2l is really blockbuilding with varying degrees of restriction - NOT just dumb pairs -and can be as advanced as you make it.


Hope this is clear and helps someone; all feedback is welcome,

Rowan.

*One great thing that can be done with EO tricks is using keyhole techniques to throw out bad edges and/or put in good edges, so you can preserve or change their orientation after a rotation. Play around with the form: (<D>) (<R, U>/<L, U>) (<D>)' and see what kind of stuff you can come up with :)
 
Last edited:

Godmil

Premium Member
Joined
Aug 12, 2010
Messages
2,196
Likes
5
Location
Aberdeen, Scotland
YouTube
Godmil
#2
Excellent! Thanks for making that guide. After someone asked about edge orientation on one of the Faz threads I thought I'd have a go at keeping the LL edges oriented during F2L - and with a lot of concentration - I ended up with a Dot OLL! So this guide is greatly appreciated.
 
Joined
Jul 23, 2008
Messages
3,375
Likes
75
WCA
2008KINN01
YouTube
RowanKinneavy
Thread starter #3
Aye well, this guide cares a lot more about f2l edges than LL edges - go over Lucas' ELS page and some VH/ZBf2l pages to get a better idea of how that sort of stuff works :)
 

Toad

Premium Member
Joined
Jul 31, 2009
Messages
1,535
Likes
0
Location
UK
WCA
2009SMIT06
YouTube
cubetoad
#6
This is most definitely the greatest little tips guide I've read on 3x3 for a long time.

I love you so much Rowan and really think this will get me so much more excited about F2L again :)
 

Godmil

Premium Member
Joined
Aug 12, 2010
Messages
2,196
Likes
5
Location
Aberdeen, Scotland
YouTube
Godmil
#7
Aye well, this guide cares a lot more about f2l edges than LL edges - go over Lucas' ELS page and some VH/ZBf2l pages to get a better idea of how that sort of stuff works :)
Yep sorry, I'd replied before I'd finished reading the OP. That was really interesting, I never thought about looking at F2L edge orientation before. I get that you may not do those kind of tricks every time in a speedsolve, but is it something you look at, just so you're subconsciously aware of the state of the cube during the solve?
 
Joined
Oct 12, 2009
Messages
62
Likes
0
#8
So how do you recognize this while solving? Do you look for edge orientation while solving the cross, or do you look for the first pair? It seems like recognizing and deciding what to do would disrupt the flow of solving unless you recognized it in preinspection after planning the cross (which sounds pretty hard).
 
Joined
Jul 24, 2008
Messages
3,812
Likes
38
Location
West Chester, PA
WCA
2008KORI02
YouTube
StachuK1992
#10
I've been finding little tidbits of sequences to do EO during F2L, so I'll try to show an example as well.

D' L' B2 D2 B' L' D' R2 D R2 U2 D2 L' R2 D2 U R F B' U B2 L2 D' F2 D2

x2 F' R' B2 F2 D F2 (6)
Here, I would scan the E-slice edges, and note that FR and BL were bad.
I try to get those done ASAP.
y U' L' U L2 F' L' F (7/13)
I now need to do a y' or y to fix the E-slice
now BL is in FR, so I'm going to get that pair over with
y U2 R U' R' x' U' R U l' (8/21)
Here, I see I can set up a quick and dirty Im case with
U' L' U L (4/25)
Then finish FR pair, and CLS it up with a mirrored I alg to avoid another rotation.

Probably doesn't help as much as Rowan't examples, but maybe a bit. :)

-statue

edit: first copy was missing a y and had an L instead of an L'. :p Fixed
 
Last edited:
Joined
Dec 27, 2008
Messages
141
Likes
0
#11
this sounds viable only to non-CNers... D: It /might/ be possible to look at orientation for a CN'er, but that will take quite a lot of practice.

BUT I AM WILLING.
 
Joined
Jul 23, 2008
Messages
3,375
Likes
75
WCA
2008KINN01
YouTube
RowanKinneavy
Thread starter #13
this sounds viable only to non-CNers... D: It /might/ be possible to look at orientation for a CN'er, but that will take quite a lot of practice.

BUT I AM WILLING.
This is funny because the only reason I made this thread is because somebody queried Feliks when he mentioned f2l EO control :)

@Ninjabob - This is like my last guide in that it's something that's useful to understand and practice, but not an 'f2l method to adopt' per se. The reason I'm making these (and have more to make) is to show people the kind of paths you go down when you're trying to optimise your f2l. By outlining my experiences with these ideas I can save people a lot of time and effort :p

As well as understanding these last two threads in a positive sense (use more tricks or reduce rotations etc), you also increase your understanding in a 'negative' sense - you can see what makes a bad f2l more easily - and make yourself more consistent.
 
Joined
Aug 14, 2007
Messages
533
Likes
3
Location
Chicago,IL
WCA
2012JAME03
YouTube
jeff081692
#19
So would it make sense to move towards an edge bias over corner bias when searching for pairs to best utilize something like this? It seems that with edge bias you can immediately see how many rotations you would need and can choose to do the rotations early or late if you took an assessment of edge orientations during the beginning of F2L. For example lets say while solving your first pair you quickly scan for the 3 remaining edges and notice that 2 of them are good edges and 1 is bad. Had you used corner bias you might end up finding the bad edge first and if you don't spot a better pair faster you end up rotating and being left with 2 bad edges so you have to rotate again while solving one of the bad edges. But if you spot 2 good edges and a bad edge you can solve the two good edges first and prepare yourself for the rotation in the final pair maybe even forcing the rotation to leave you with the last pair in the FR slot to help with OLS or something. Or are top cubers who have corner bias during F2L able to see more or less the flow of the solve because they can find corner and edge pairs so quickly that they can make a decision between rotating now or later just by glancing at all the pairs.

I kind of understand how you can have at most one rotation per F2L if you solve all the good edge F2L pairs first then rotate to make all the bad edges good and then solve them (assuming none of the algorithms flipped the orientation of an edge).

Old thread but I have been wondering how much thought I should give to this while doing solves as I think I could develop the ability to distinguish good and bad edges quickly especially after trying Roux and maybe during inspection it could even be possible to influence the edges to make F2L as smooth as possible but I am not at the level to know if that would even be worth it or not.
 
Top