Why You Should Use EJF2L (Video)

[JLarsen]

If you go to the actual video and look there are 2 video responses from Erik
Johnson; one for the algs, and one for examples of where and how he uses it.

 

[soccerking813]

Not sure if this is just me, but it says the video is no longer available.

 

[Samlambert]

Not sure if this is just me, but it says the video is no longer available.

Its just you.

On-topic: Interesting but I don't think its really worth it.

 

[mcciff2112]

that sounds really interesting. might just have to check that out. thanks

 

[JLarsen]

Not sure if this is just me, but it says the video is no longer available.

Its just you.

On-topic: Interesting but I don't think its really worth it.

Not worth what? Switching to Petrus? Trying a method besides Fridrich? Now if your like sub 17 with Fridrich already, and you just don't want to change methods, then I understand. I would use Roux, but I've been doing Petrus for 8 months.

 

[qqwref]

I find it absolutely hilarious that you say Fridrich requires memorization of an "absurd amount of algs" (21+57 = 78, right?) but don't mention that at all for CLS (which is 104 algs)

Anyway EJF2L is 7+8+8 = 23 OCLL cases, right? I'm counting the normal F2L pair cases because I assume that if you had an easy way to insert the normal F2L pair you'd just do that and then use normal OCLL. I suppose you'd also want to learn the mirrored versions of the I and Im cases (for the RB slot) so that would be 39 cases total, which is getting up there. Cool method though, I imagine it would save some significant time if you use Petrus. I don't think it would save any time if you use Fridrich, though, and I'm not about to switch to Petrus F2L since I'm terrible at blockbuilding.

Would it be at all accurate to explain the EJF2L last pair insertion as "pair up and insert the corner and edge like you would for a normal corner/edge pair, except don't pay attention to the orientation of the corner"?

 

[JLarsen]

Yup you could explain it like that. Although suppose you got a twisted 2x2x2 skip(yes it's happened to me plenty of times), then there would be no inserting of pairs.
Now when I said something about the memo, I was talking about how normal MGLS takes 34 ELS algs, and 104 CLS algs, which is absurd. With petrus you only need 16 algs, thats it. I'll give you the links;

http://cube.garron.us/MGLS/I.htm

Now I do know 7 olls' all pll's, and those 16 algs for EJ. which makes 7+16+21 is 44 yes? So my total is less than the 57 oll's alone used in Fridrich.

Edit;When I explain MGLS i note 104 cases for full CLS....=\

Edit 2; also, I really do agree that with Fridrich, it can't take your times down anymore than learning VH alone, and leaving out CLS altogether.

 

[Kyle™]

There's lots of stuff out there that seems really good, but turns out to be too much work. For example I found these on Dan Harris' site I think, and fixed and modified them all http://www.freewebs.com/cubesolution/COllalgs.rtf

Still no one uses it. Too much work.

More algorithms= faster solving times. Just how the world works.

 

[EmersonHerrmann]

There's lots of stuff out there that seems really good, but turns out to be too much work. For example I found these on Dan Harris' site I think, and fixed and modified them all http://www.freewebs.com/cubesolution/COllalgs.rtf

Still no one uses it. Too much work.

More algorithms= faster solving times. Just how the world works.

Not necessarily...I know people who know all the fridrich algs, and average like...19. While Erik Johnson knows like...44 algs and averages sub-14. So no, that's not really true.

 

[Kyle™]

I only know maybe 5 fridrich algs and can get 18 avg. I am really poor at that method... so whoever you know that averages 19 just needs more practice...It has nothing to do with the number of algorithms they know.

Also, there is a difference between getting one average of sub-14, and averaging sub-14 consistently. Erik knows as well as any petrus user, that it is difficult to average low because of the majority of difficult and tedious cases.

 

[EmersonHerrmann]

I only know maybe 5 fridrich algs and can get 18 avg. I am really poor at that method... so whoever you know that averages 19 just needs more practice...It has nothing to do with the number of algorithms they know.

Also, there is a difference between getting one average of sub-14, and averaging sub-14 consistently. Erik knows as well as any petrus user, that it is difficult to average low because of the majority of difficult and tedious cases.

You know he is consistently sub-14, right?

And when you say "needs more practice", you're disagreeing with what you said earlier. You said earlier that the more algs you use the faster are.

Nevermind, it's pointless arguing over this heheh

 

[CanadianPires]

Seems interesting to me but its not sounding like much of an advantage, here's my train of thought:

Petrus user, builds 2x2x3 block. EO. Then builds last 2x2x1.

Now, normal petrus user builds last CE pair and is left with 7 OCLL cases.

EJF2L user builds last CE pair w/o regard to corner orientation and then is left with 16 CLS cases.

.. How is this a great advantage, over twice as many algs for a step (normal CE pair vs. EJ pair) that doesn't seem to be that much easier or quicker to recognize.

I know theres also the case of having an EJ pair in the 2x2x3 block, but still?

 

[EmersonHerrmann]

I'm learning these algs...for no reason

 

[pinoycuber]

Erik's idea is very good.. im still on study about it.. but this video helps thanks..

 

[JLarsen]

Seems interesting to me but its not sounding like much of an advantage, here's my train of thought:

Petrus user, builds 2x2x3 block. EO. Then builds last 2x2x1.

Now, normal petrus user builds last CE pair and is left with 7 OCLL cases.

EJF2L user builds last CE pair w/o regard to corner orientation and then is left with 16 CLS cases.

.. How is this a great advantage, over twice as many algs for a step (normal CE pair vs. EJ pair) that doesn't seem to be that much easier or quicker to recognize.

I know theres also the case of having an EJ pair in the 2x2x3 block, but still?

Have you ever tried blockbuilding? It can make things incredibly convenient, this does not in any way just apply to the last slot, or just the 2x2x3. It can be done anywhere. Meaning you can make a twisted 2x2x2, or in either "slot" during step 4. My favorite case happens to be the "EJ Square"=]. Say you get to step 4 and you have a 1x2x2 with a twisted corner. You can save like 2 seconds just by throwing it in there, and this does happen quite a lot. Another thing; a couple seconds means everything once you hit sub 20, and its the difference between a good and a bad average for people like Erik. And as far as recognition, its just like learning how to blockbuild, you get an eye for it. Lastly, I don't think you realize that EJF2L is not used in every solve. It is used when it is appropriate, meaning the solution is faster, easier, or just overall nicer. It's a blockbuilding technique, merely an addition to Petrus, and the time it takes to memo 16 2 gen algs, is certainly worth the time for this convenience in my solves. Besides, I doubt you use Petrus as your main method ;]

Maybe that clears things up a bit?

 

[CanadianPires]

Seems interesting to me but its not sounding like much of an advantage, here's my train of thought:

Petrus user, builds 2x2x3 block. EO. Then builds last 2x2x1.

Now, normal petrus user builds last CE pair and is left with 7 OCLL cases.

EJF2L user builds last CE pair w/o regard to corner orientation and then is left with 16 CLS cases.

.. How is this a great advantage, over twice as many algs for a step (normal CE pair vs. EJ pair) that doesn't seem to be that much easier or quicker to recognize.

I know theres also the case of having an EJ pair in the 2x2x3 block, but still?

Have you ever tried blockbuilding? It can make things incredibly convenient, this does not in any way just apply to the last slot, or just the 2x2x3. It can be done anywhere. Meaning you can make a twisted 2x2x2, or in either "slot" during step 4. My favorite case happens to be the "EJ Square"=]. Say you get to step 4 and you have a 1x2x2 with a twisted corner. You can save like 2 seconds just by throwing it in there, and this does happen quite a lot. Another thing; a couple seconds means everything once you hit sub 20, and its the difference between a good and a bad average for people like Erik. And as far as recognition, its just like learning how to blockbuild, you get an eye for it. Lastly, I don't think you realize that EJF2L is not used in every solve. It is used when it is appropriate, meaning the solution is faster, easier, or just overall nicer. It's a blockbuilding technique, merely an addition to Petrus, and the time it takes to memo 16 2 gen algs, is certainly worth the time for this convenience in my solves. Besides, I doubt you use Petrus as your main method ;]

Maybe that clears things up a bit?

Fair enough, but watching the video, one gets the idea that he is only referring to the last slot. In which case, my comparison of a normal Petrus last slot and an EJ last slot still holds. I agree though, that if you insert it at anytime in the F2L it can be advantageous. Maybe not hugely, but for advanced petrus users, if you don't mind memo'ing 16 extra algs, go for it.

 

[JLarsen]

=]

I'm sorry I wasn't very clear on that in my video, Ill add an annotation in there or something.