after messing around with the method for about a month, here are some of my thoughts, as an average sub 1 minute 4x4 solver L2E: everyone should learn it, no doubt L3E - all on U face: I've hand genned some more algs for this type of cases, the recognition is fine, and not all of them have an unique alg, some are just another with some setup moves if you know what I'm saying. Definitely worth learning and experimenting L3E - 1 on E slice: Doing block comms and RUM comms is fine, the recognition is quite fast after a while actually L4E - all on U face: I think only cases that are 2e2e type stuff worth generating direct algs, reducing some other more awkward cases into 2e might not be as speed-efficient, instead just solve one edge and do L3E. L4E - 1 on E slice: it depends on what I was geting, I would do a setup move into some L4E all on U face cases, whether reduce it into L3E. L5E: I don't think recognition for pure L5E worths the effort, but 2e3e/3e2e is ok.
Only a number cubers have Jayden's knowledge and experience. But when it comes to teaching less experienced cubers like myself - i can tell you from experience that his teaching skills are even better :)
Always wanted to experiment with something like this. I've been trying to use block comms more and more for l3e and I think this "method" could genuinely be very effective (some cases are a bit iffy and can probably just be chain-paired). I'll have to look into it some more and see if I can make it work
I'd be curious to see what you can find with some of these ideas! At the moment I think the main things holding this "method" back are the non block comm 3 cycles with 1 in E and two in U, and also the 4E scenario feels like it's not truly optimal
for the 2e2e I feel like 1 move setup to one of those 3 algs(H perm and Etter olls) is worth using. that's R' into alg or R' with U into alg. if the last equator edge was in FR you could've used either R of F setups, potentially if this way of edge solving is the future, then we might rethink our first 3 edges of 3-2-3 to have the last one in front for slightly better solutions on average. besides that you can force one of those 3 2e2e cases with a trigger almost all of the time(maybe even all of the time), but that would be too hard to recognize quickly. maybe if we had 1 more fast 2e2e cases then we would only need to recognize the fact that we have a 2e2e, but not the specific case, then start putting 2 edges we already know the state of in some position (always opposing each other or oriented) while leaving the other 2 edges on top, and while doing those moves(a trigger) we can recognize the other 2 edges and know what alg to use. actually if we can put only 2 edges on LL opposite of each other while forcing their orientation to be like in H perm 2e2e, then we can already use this strat, because the only cases you can get are H perm and Rw U Rw' R U R'... alg My other thought was to use the H perm 2e2e to create 3edge cases out of 5e or 3e2e or 4e. you would need an in-depth recog strat, but maybe it's possible, though the swiftness of 2e alg and it's ability to lookahead into the 3e that you will have left, probably makes this idea useless. Interesting aspect of this video's idea is that F2L lookahead is easier, because you have FL and FR slots almost unused and the shorter 3e algs change the LL in a really predictable manner. theoretically you can do an L' before one of the algs to use f2l pieces stuck in the FL slot. There's an idea for the first 3 edges of 3-2-3 which might be worth looking into in combination with this L5E idea(the video). instead of slicing away the edge, then finding it's 2nd half and doing that with the next edge. You can do Uw' while looking at FRd wing and try to insert it into BR, and while you are inserting or looking for it you can notice which wing is in the FRu after the initial Uw' and try to find it and insert into FL. Thus you almost simultaneously are solving 2 edges. With better lookahead you could also search and insert those 2 edges in any order and you can theoretically also notice the wing that went into RBu after the first Uw' and if nothing good is apparent for FR edge, you can solve normally(for a Uw' solvers). Another even harder idea for this is to notice the BLd edge while finishing your 4th cross edge, or even while finishing the centers because the edges in-between 2 solved centers don't usually change. and then you can do Uw while recognizing FLu wing that goes into BLu and now you can do the initial idea of inserting 2 edges into FL and BR, but this time you don't have a blind spot in BL and afterwards you will have unsolved UF instead of BR
I might hate 4x4 less now. Been doing a “Dylan” version of 3-2-3 just to avoid a lot of bad cases that come up and make look ahead smoother… might see about incorporating some of these.
I might try to gen the algs, but probably not for a couple of weeks. Question: how viable do you think knowing all the cases would be (all 2-5 cycles)? I might be wrong, but there's like 100-150 cases I think. Obviously recognition would be bad, but if you get to solve everything in one step, maybe its worth it? I know some people know all the 4-cycles if they are on top, so this would be kinda the same just taken a bit further. It would be a lot of algs, but probably comparable to EG, so not an impossible amount to learn.
I'd be curious to see what you can gen I feel like 5 cycles would almost definitely be unviable in actual solves due to the amount of time it already takes to differentiate between 5E and 2E3E, let alone recognizing beyond that. They're probably still worth genning as a proof of concept or for fun, but I feel that algs for full 2-4E would have a much better chance of being something worthwhile for fast people I've already found a few other algs by hand btw like Rw D Rw' U2 Rw D' U R' U' R Rw'. I also know some intuitive tricks exist like slice R' U R' F R2 F' (or inverse) slice back to solve 4E, but these feel incompatible with 3-2-3 due to needing to set up multiple edges in the E slice instead of the U layer
I def hand genned these but for some reason I forgot about them or didn't like them for some reason (I think it's because they don't transfer the best when mirrored across F/B) They're definitely viable candidate algs though, I think r' F R' F' Rw and the F/B mirror might be worth just because they start the same way as the "good" versions of those cases
perhaps by analogy with the word parity, it would appear this jayden mcneill fellow has spelled the word "pairing" as "paring [sic]". however, this spelling would actually refer to the gerund form of the verb "to pare", meaning to trim, and is presumably not what is meant by the title of this video. in light of this, i humbly advise a rectification of this error, so as to better reflect on the professionalism of this channel.
i love it when jayden mcneill gets bored
We need him to get bored more
Yoo charlie
You getting bored leads to me getting the opposite of bored. Feel free to get bored more often.
Timestamps:
Setups to L2E or L3E:
2:06 2E2E
4:39 4E
9:39 2E3E
11:38 3E2E
13:08 5E
L2E and L3E algs:
14:07 good/oriented
15:17 bad/oriented
15:40 good/mismatch
18:08 bad/mismatch
20:34 U+pair
21:41 U+mismatch
22:41 U+ls
thanks
4x4 FMC has opened his mind
Any 4x4 can be solved in 55 moves or less (OBTM). I wonder if he can pull off a sub-55...
after messing around with the method for about a month, here are some of my thoughts, as an average sub 1 minute 4x4 solver
L2E: everyone should learn it, no doubt
L3E - all on U face: I've hand genned some more algs for this type of cases, the recognition is fine, and not all of them have an unique alg, some are just another with some setup moves if you know what I'm saying. Definitely worth learning and experimenting
L3E - 1 on E slice: Doing block comms and RUM comms is fine, the recognition is quite fast after a while actually
L4E - all on U face: I think only cases that are 2e2e type stuff worth generating direct algs, reducing some other more awkward cases into 2e might not be as speed-efficient, instead just solve one edge and do L3E.
L4E - 1 on E slice: it depends on what I was geting, I would do a setup move into some L4E all on U face cases, whether reduce it into L3E.
L5E: I don't think recognition for pure L5E worths the effort, but 2e3e/3e2e is ok.
Only a number cubers have Jayden's knowledge and experience. But when it comes to teaching less experienced cubers like myself - i can tell you from experience that his teaching skills are even better :)
I’m always glad when Jay posts these types of videos
As someone who mains the YJ Zhilong Mini 4x4, your cubes look absolutely massive, great video!
Always wanted to experiment with something like this. I've been trying to use block comms more and more for l3e and I think this "method" could genuinely be very effective (some cases are a bit iffy and can probably just be chain-paired). I'll have to look into it some more and see if I can make it work
I'd be curious to see what you can find with some of these ideas! At the moment I think the main things holding this "method" back are the non block comm 3 cycles with 1 in E and two in U, and also the 4E scenario feels like it's not truly optimal
for the 2e2e I feel like 1 move setup to one of those 3 algs(H perm and Etter olls) is worth using. that's R' into alg or R' with U into alg.
if the last equator edge was in FR you could've used either R of F setups, potentially if this way of edge solving is the future, then we might rethink our first 3 edges of 3-2-3 to have the last one in front for slightly better solutions on average.
besides that you can force one of those 3 2e2e cases with a trigger almost all of the time(maybe even all of the time), but that would be too hard to recognize quickly.
maybe if we had 1 more fast 2e2e cases then we would only need to recognize the fact that we have a 2e2e, but not the specific case, then start putting 2 edges we already know the state of in some position (always opposing each other or oriented) while leaving the other 2 edges on top, and while doing those moves(a trigger) we can recognize the other 2 edges and know what alg to use.
actually if we can put only 2 edges on LL opposite of each other while forcing their orientation to be like in H perm 2e2e, then we can already use this strat, because the only cases you can get are H perm and Rw U Rw' R U R'... alg
My other thought was to use the H perm 2e2e to create 3edge cases out of 5e or 3e2e or 4e. you would need an in-depth recog strat, but maybe it's possible, though the swiftness of 2e alg and it's ability to lookahead into the 3e that you will have left, probably makes this idea useless.
Interesting aspect of this video's idea is that F2L lookahead is easier, because you have FL and FR slots almost unused and the shorter 3e algs change the LL in a really predictable manner. theoretically you can do an L' before one of the algs to use f2l pieces stuck in the FL slot.
There's an idea for the first 3 edges of 3-2-3 which might be worth looking into in combination with this L5E idea(the video). instead of slicing away the edge, then finding it's 2nd half and doing that with the next edge. You can do Uw' while looking at FRd wing and try to insert it into BR, and while you are inserting or looking for it you can notice which wing is in the FRu after the initial Uw' and try to find it and insert into FL. Thus you almost simultaneously are solving 2 edges. With better lookahead you could also search and insert those 2 edges in any order and you can theoretically also notice the wing that went into RBu after the first Uw' and if nothing good is apparent for FR edge, you can solve normally(for a Uw' solvers).
Another even harder idea for this is to notice the BLd edge while finishing your 4th cross edge, or even while finishing the centers because the edges in-between 2 solved centers don't usually change. and then you can do Uw while recognizing FLu wing that goes into BLu and now you can do the initial idea of inserting 2 edges into FL and BR, but this time you don't have a blind spot in BL and afterwards you will have unsolved UF instead of BR
The Knowledge King has returned to bless us with a fraction of his power
I might hate 4x4 less now.
Been doing a “Dylan” version of 3-2-3 just to avoid a lot of bad cases that come up and make look ahead smoother… might see about incorporating some of these.
I might try to gen the algs, but probably not for a couple of weeks.
Question: how viable do you think knowing all the cases would be (all 2-5 cycles)? I might be wrong, but there's like 100-150 cases I think. Obviously recognition would be bad, but if you get to solve everything in one step, maybe its worth it? I know some people know all the 4-cycles if they are on top, so this would be kinda the same just taken a bit further. It would be a lot of algs, but probably comparable to EG, so not an impossible amount to learn.
I'd be curious to see what you can gen
I feel like 5 cycles would almost definitely be unviable in actual solves due to the amount of time it already takes to differentiate between 5E and 2E3E, let alone recognizing beyond that. They're probably still worth genning as a proof of concept or for fun, but I feel that algs for full 2-4E would have a much better chance of being something worthwhile for fast people
I've already found a few other algs by hand btw like Rw D Rw' U2 Rw D' U R' U' R Rw'. I also know some intuitive tricks exist like slice R' U R' F R2 F' (or inverse) slice back to solve 4E, but these feel incompatible with 3-2-3 due to needing to set up multiple edges in the E slice instead of the U layer
Hey Jay, it's been a while, how's it going? Awesome video by the way :))
Omg he actually posted it!
GOLD
@@vspython3385 Blizzy!
@@Anthony_Jiang Been a while, hasn't it-
Yo do you know csp for squan?
missed you jay
This kinda gives me K4 F3L vibes
i like a lot
for the opp pair + misorient @ 21:11 you can also do R U R' U' r U R U' Rw' from that angle if you find the F moves too awkward
also at 8:37 you did the wrong alg which worked out cause it paired up the other wing in that E layer pair lmao
At this rate you must have invented new methods in your dreams
Jayden McNeill
Next make a video on edging methods please
19:22 Rw' F R F' r. No need to do U2 beforehand.
Also 21:22 r' F R' F' Rw.
I def hand genned these but for some reason I forgot about them or didn't like them for some reason (I think it's because they don't transfer the best when mirrored across F/B)
They're definitely viable candidate algs though, I think r' F R' F' Rw and the F/B mirror might be worth just because they start the same way as the "good" versions of those cases
Even better for 21:22 is Rw U R' U' r'.
Black magic in cubing frfr
holy shit i do that all the time
hai jay :3
hi fesh
perhaps by analogy with the word parity, it would appear this jayden mcneill fellow has spelled the word "pairing" as "paring [sic]". however, this spelling would actually refer to the gerund form of the verb "to pare", meaning to trim, and is presumably not what is meant by the title of this video. in light of this, i humbly advise a rectification of this error, so as to better reflect on the professionalism of this channel.
i eat 4x4s
me too what's your favorite flavor
@@notanonymousperson i like mine with some flavoured lube, wbu?
@@notanonymouspersonjw what
Bored Jay is bored.
He must be to like a comment that was a meme in 2009.
Seriously, haven't you left the house or talked to anyone in 15 years?
is it just me or im dumb
me too