The New Method / Substep / Concept Idea Thread

[PiKeeper]

I no longer trust your judgement

I thought everyone now agreed that Mehta was objectively worse Petrus?

 

[IsThatA4x4]

I thought everyone now agreed that Mehta was objectively worse Petrus?

Yup we found out that mehta ends up worse in many of the ways it tried to be better...
It's a cool method but just not objectively among the top.

 

[CornerTwisted]

People are just not good at Mehta right now. Looking at some statistics, it could be a godly method if somebody decided to speedsolve with it. The first few steps are intuitive and therefore require a lot of look ahead, and looking ahead to EO is really all you need. The rest of the method is highly ergonomic-friendly algorithms, making the solving experience optimised and fast.

 

[Timona]

Yup we found out that mehta ends up worse in many of the ways it tried to be better...
It's a cool method but just not objectively among the top.

People haven't committed enough time to learning it and becoming world-class with it as with CFOP or Roux or even ZZ. Those methods have been around for more than 7 years, at least. Mehta is not even 3 years old at this point. If the method had been around for longer, it would probably be a different story. It's just gonna take time.

The end of the method is mostly algorithmic, so people can start learning how to cancel between steps and stuff like that.

 

[OreKehStrah]

People haven't committed enough time to learning it and becoming world-class with it as with CFOP or Roux or even ZZ. Those methods have been around for more than 7 years, at least. Mehta is not even 3 years old at this point. If the method had been around for longer, it would probably be a different story. It's just gonna take time.

The end of the method is mostly algorithmic, so people can start learning how to cancel between steps and stuff like that.

Mehta-TDR (the best version) was also invented by a French person in 2017 and then abandoned because it sucked, only to be rediscovered and again be found not worth using.

 

[PiKeeper]

People haven't committed enough time to learning it and becoming world-class with it as with CFOP or Roux or even ZZ. Those methods have been around for more than 7 years, at least. Mehta is not even 3 years old at this point. If the method had been around for longer, it would probably be a different story. It's just gonna take time

It's not that Mehta is new and that's why it's bad. It's that it has been proven to be worse than Petrus/APB. Athefre did a comparison between Mehta, Petrus, APB, and maybe some other variants and the results conclusively showed Mehta to be worse. Most of the people who initially used Mehta switched methods because of this. People need to accept that Mehta is not ever going to surpass Petrus/APB.

 

[Deleted member 36445]

It's not that Mehta is new and that's why it's bad. It's that it has been proven to be worse than Petrus/APB. Athefre did a comparison between Mehta, Petrus, APB, and maybe some other variants and the results conclusively showed Mehta to be worse. Most of the people who initially used Mehta switched methods because of this. People need to accept that Mehta is not ever going to surpass Petrus/APB.

links on petrus/apb please?

 

[Timona]

links on petrus/apb please?

APB on the Wiki. Main Site.

Petrus.

 

[Silky]

isn't this just a slightly worse version of petrus?

Every method is a worse version of Petrus

 

[Jouda_209]

Hello my friends,
I was just wondering around, trying out stuff with ZZ and i thought about something. You know there is a type of ZZ called portico where instead of EOline you just solve one piece of EOline and ignore the other. Well as many of you know almost nobody now uses EOline, so i thought what if i do EOcross but ignore the DF edge? And than for last slot added winter variation of VLS. Turns out you are left with L5EP and CP. And i thought ok when i solve L5EP there are only 5 cases you can get after and those are Aa perm, Ab perm, H perm, E perm and solved state. But turns out the algs i found said there are 16 cases for L5EP, but it didnt include algs that don't have solved CP. So i did some diging and turns out, I cant find any of the algs I was looking for. If someone does know where to find these algs are, or has these alg please send them to me, i will be very grateful.

 

[PiKeeper]

L5EP doesn't involve the corners at all, so the algs don't change

 

[Jouda_209]

L5EP doesn't involve the corners at all, so the algs don't change

Well it doesnt, but you can get some sort of parity, becquse you can have just 2 edges and 2 corners swapped.

 

[PiKeeper]

Well it doesnt, but you can get some sort of parity, becquse you can have just 2 edges and 2 corners swapped.

I guess you could gen them yourself with trangium's batch solver, or you could just solve corners and then do L5EP like everyone else.

 

[LukasCubes]

kinda stupid 3x3x4 method i use

hold cube vertically

corners (dont care how you solve them)
bottom 4 edges (dont care how you solve)
top 4 edges in one look

rotate so the U/D layers are now L/R

Solve and finish one center and put that on D
solve the edges next to said center whilst solving the center above that, one at a time, by now, all centers are done
L4E, ALL IN ONE ALG

woohoo you did my corners first method

 

[CornerTwisted]

woww wow

 

[Xatu]

i came up with a method called ZZ-ZPL2
step 1: EO222
step 2: F2L-1
step 4: insert last corner while making 2x2x1 block on backright. And CP
step 5: solve with 18 algorithms

i invented a method called ZZ-ZPL2 which is a ZZ variant.
i'm here to say that it is not only very efficient, has a pretty low movecount, RUL gen, and only has 18 algs,
in theory, it’s very good.

 

[PiKeeper]

i came up with a method called ZZ-ZPL2
step 1: EO222
step 2: F2L-1
step 3: CP
step 4: insert last corner while making 2x2x1 block on backright.
step 5: solve with 18 algorithms

Is this really what you spent all that time harping about as one of the best methods? It's worse zz-d, which is itself worse than zz-a.

 

[Xatu]

Is this really what you spent all that time harping about as one of the best methods? It's worse zz-d, which is itself worse than zz-a.

stop that. its a good method.

Hello my friends,
I was just wondering around, trying out stuff with ZZ and i thought about something. You know there is a type of ZZ called portico where instead of EOline you just solve one piece of EOline and ignore the other. Well as many of you know almost nobody now uses EOline, so i thought what if i do EOcross but ignore the DF edge? And than for last slot added winter variation of VLS. Turns out you are left with L5EP and CP. And i thought ok when i solve L5EP there are only 5 cases you can get after and those are Aa perm, Ab perm, H perm, E perm and solved state. But turns out the algs i found said there are 16 cases for L5EP, but it didnt include algs that don't have solved CP. So i did some diging and turns out, I cant find any of the algs I was looking for. If someone does know where to find these algs are, or has these alg please send them to me, i will be very grateful.

yaeh i invented a better zz variant called ZZ-ZPL2 That ignores one line edge

 

[PiKeeper]

stop that. its a good method.

You can't just say that with no evidence. You can start by proving why it's better than zz-d which is eocross, f2l-1, solve pair and cp, 2gll.
Edit: Actually, you can just compare it to to zz-tripod. That's just your method but better since it doesn't involve cp.

 

[IsThatA4x4]

yaeh i invented a better zz variant called ZZ-ZPL2 That ignores one line edge

I don't mean to be rude (I don't like doing that), but you really need to accept other peoples' criticism of your method.
If you think it's a good method, you should provide arguments as to why it is a good method, and you should take into account why other people think it is good/bad.
I'll admit, the actual amount of valid criticism other than "this is bad/worse version of x" here is very minimal, so I'll try to explain what I think about the method here, and maybe some others will do the same.

Your first steps are fine, reaching EOF2L-1 is normally a good starting point for a ZZ variant, however EO222 is much more reminiscent of petrus than ZZ, which would solve an EOline/arrow/cross of some sort as its first step. Just doing EO with a block at the start is not really enough to call it a ZZ variant, but it doesn't matter much about how good the method actually is.

For step 3, there are a couple issues. Firstly, CP recognition is not the best (although it's not that bad at LSLL), but the main issue is doing CP followed immediately by solving the corners. It would be better to do something at the same time as CP to make the method more move efficient and require less potential pausing, for example, using a technique like conjugated CLL to solve all the corners in one go, or doing CP and the block at the same time, and then solving the corners next.

Finally, for the last step, I don't quite know where you're going, as solving F2L-1, a 2x2x1 block on top, the corners, and EO leaves you with 3 edges to solve, which is only 2 algorithms.
This is a pretty good sign that you could sacrifice something from one of your previous steps, as ending with one of 2 algs shows that you have reduced the cube's state a little too far for less benefit. If, for example, you only solve 1x1x2 block on top, you would have 4 edges left, although that's still not many cases.
If you just did the corners and then finished, you'd have ZZ-zipper L5E (around 20 algs I think), which is better.
However, I think you might possibly mean something else with your steps, I'm not quite sure, but judging from what else you've said about missing one line edge, you could mean missing out an edge in the bottom layer too, in which case having that + an E layer edge unsolved + top edges will yield algs that are mediocre at best.

Right, there's my criticism, it was lengthier than I expected, but I really wanted to try and explain what's holding you back here.
Please take others' opinions into account, otherwise people will not think much of what you're saying if it's not backed up.

Hope I've helped in some way