OH Method Debate

[antonio19kamh3]

thing with lse vs 2gll is that with lse, you don't really have any recognition issues and can just do it pauseless after cmll, but with 2gll, you need to pause to recognise, which wastes quite a lot of time in comparison

 

[slowhandzboi]

I think YruRU is the best OH method.

YruRU is actually Briggs method

 

[brr_ya]

How does CP affect the solve?

 

[AlphaCuber is awesome]

How does CP affect the solve?

It allows it to be fully 2gen after EO stripe and reduces alg count

 

[mukerflap]

How does CP affect the solve?

it means you have a 2 gen last layer.

 

[HaHaHaHeeHeeHee]

Speaking of Why roux, how the heck to you fingertrick M (not M' that's easy) it makes no sense that people can do M that fast. So my question instead is How roux?

lol its pretty easy u put the cube on the dr edge on the table at 45 degrees and u flick the f sticker of the df edge with pinky finger
and also I think that Roux is better in terms of proportionality to ur 2h times
cuz I average like 20 with cfop and about 40 oh with cfop
and I average like 17 with Roux and average about 25-26 oh with roux

the main thing that I think makes me slow for cfop oh is the L and L' moves in f2l compared to the R and r moves of f2b
so out of cfop and roux I say roux is better

cant speak about vaayrooroo cuz I dont know it but I hear that its a kinda bot method so im not gonna look into it

 

[PizzaCuber]

Why does this have 8 pages already and it was posted yesterday...

 

[ep2]

Why does this have 8 pages already and it was posted yesterday...

Because there's another 4 pages of discussion in another thread. It should be about 12.

 

[HaHaHaHeeHeeHee]

GJ! It took you 30+ posts to make one post proving your point.

If Roux was a TH killer, CFOP would've died 16 years ago. Therefore, Roux is trash.

I think the reason that CFOP is still the main method is for a few reasons
1. it easily transfers from LBL, which is why I started out with it
the next one I think is partially my problem
2. Roux has extremely and I mean extremely few resources- I only figured out that there was a roux method from j perms overview video
which led me to Kians videos... so I think that this is the reason that less people use roux compared to cfop
3. idk why, but ppl seem to think that blockbuilding is hard. I learnt Roux last december and averaged like 2 minutes per solve and that was using pll algs to skip ul ur and oll to skip eo but in like 2-3 months i got to sub 25 so its not hard at all


and also Sean Villanueva is 3rd WR for 3x3 avg and Kian Mansour is 3rd WR for OH avg
also before Kian are Max and Patrick, but then you have to look at the difference in 3x3 between Max and Patrick and Kian vs the difference in Oh, which is a very small fraction of a second

 

[brododragon]

I think the reason that CFOP is still the main method is for a few reasons
1. it easily transfers from LBL, which is why I started out with it
the next one I think is partially my problem
2. Roux has extremely and I mean extremely few resources- I only figured out that there was a roux method from j perms overview video
which led me to Kians videos... so I think that this is the reason that less people use roux compared to cfop
3. idk why, but ppl seem to think that blockbuilding is hard. I learnt Roux last december and averaged like 2 minutes per solve and that was using pll algs to skip ul ur and oll to skip eo but in like 2-3 months i got to sub 25 so its not hard at all


and also Sean Villanueva is 3rd WR for 3x3 avg and Kian Mansour is 3rd WR for OH avg
also before Kian are Max and Patrick, but then you have to look at the difference in 3x3 between Max and Patrick and Kian vs the difference in Oh, which is a very small fraction of a second

I was making a point of how methods don't just become the greatest instantly.

 

[Etotheipi]

I think the best OH method is going to be one specifically designed for OH. None of the current main OH methods had this in mind when they were developed as far as I'm aware of, but YruRU/Briggs is a step in that direction, though they may not have surpassed 2H methods yet for OH. but if OH methods continue to get more attention and development then I do think better OH methods will be found that beat standard 2H methods.

 

[RedstoneTim]

I think the best OH method is going to be one specifically designed for OH. None of the current main OH methods had this in mind when they were developed as far as I'm aware of, but YruRU/Briggs is a step in that direction, though they may not have surpassed 2H methods yet for OH. but if OH methods continue to get more attention and development then I do think better OH methods will be found that beat standard 2H methods.

I can't agree with you here at all.

Firstly, a lot of people care about OH and are actively trying to improve at it. That's mainly because OH is not that hard to learn when you already know normal 3x3 (it's mainly getting used to new fingertricks and slightly altering your solving style, for example more EO influencing or fixed BL edge) and because there's more room for method improvement than for 2H since the currently dominating method, CFOP, is not nearly the best OH method of the big four. By no means does this mean that CFOP is a bad OH method - it's actually pretty good as evidenced by the world records and rankings -, but Roux is objectively superior. This means that OH already gets a lot of attention: especially Rouxers are constantly developing their method to be better for OH.

My second point is that a method doesn't have to be designed specifically for something to be good for it. Of course if you want to create a good OH method, there's no reason to invent a method for 2H and just hope it's good for OH as well. But that doesn't mean a method designed specifically for 2H can't be very good for OH after some adjustments.
Take Roux as an example: no one a decade ago thought that Roux had lots of potential for OH - until Thom Barlow discovered table abuse and found out that this would revolutionize OH Roux. That in my opinion is the sole reason that Roux is as good for OH as it's now, so I don't see why this couldn't happen with any other 2H method. It's great that people are creating methods specifically for OH, but you still have to keep in mind that most good ideas were initially created with a totally different purpose. Sometimes, a weird thing you had only intended for your architecture students might just become one of the best selling toys in the world - even if you wouldn't have ever expected that to happen when you invented it.

 

[Etotheipi]

I can't agree with you here at all.

Firstly, a lot of people care about OH and are actively trying to improve at it. That's mainly because OH is not that hard to learn when you already know normal 3x3 (it's mainly getting used to new fingertricks and slightly altering your solving style, for example more EO influencing or fixed BL edge) and because there's more room for method improvement than for 2H since the currently dominating method, CFOP, is not nearly the best OH method of the big four. By no means does this mean that CFOP is a bad OH method - it's actually pretty good as evidenced by the world records and rankings -, but Roux is objectively superior. This means that OH already gets a lot of attention: especially Rouxers are constantly developing their method to be better for OH.

My second point is that a method doesn't have to be designed specifically for something to be good for it. Of course if you want to create a good OH method, there's no reason to invent a method for 2H and just hope it's good for OH as well. But that doesn't mean a method designed specifically for 2H can't be very good for OH after some adjustments.
Take Roux as an example: no one a decade ago thought that Roux had lots of potential for OH - until Thom Barlow discovered table abuse and found out that this would revolutionize OH Roux. That in my opinion is the sole reason that Roux is as good for OH as it's now, so I don't see why this couldn't happen with any other 2H method. It's great that people are creating methods specifically for OH, but you still have to keep in mind that most good ideas were initially created with a totally different purpose. Sometimes, a weird thing you had only intended for your architecture students might just become one of the best selling toys in the world - even if you wouldn't have ever expected that to happen when you invented it.

I never said that 2H methods are bad for OH at all, I was just saying that methods that are designed specifically for OH can have more potential than methods that were designed for 2H and just happened to be good for OH, and as such got optimized for that. Though maybe 2H methods are the best, but I doubt that.

 

[maticuber]

In general good 2H methods have the lowest move count possible that's also easy to recognize, and guess what, that's also pretty good for OH. You obviously want to optimize the moves for OH, but F2L has always been good for OH, and LL variants have been developed and improved over the years.

10 years ago people were doing OH F2L at similar speeds as today, but LL was not that developed and usually times were decided by your LL. With the development of OH oriented COLL, more edge control algs and ZBLL OH has improved a lot. Obviously roux is better if you consider table abuse because it has less moves, and people didn't do that before not because it wasn't discovered, but because the hardware at the time wasn't as good.

IMO zz is way better than cfop for OH just because you do all your EO first and it really feels like a method developed for OH, and afaik it was developed for 2H.

 

[RedstoneTim]

I never said that 2H methods are bad for OH at all, I was just saying that methods that are designed specifically for OH can have more potential than methods that were designed for 2H and just happened to be good for OH, and as such got optimized for that. Though maybe 2H methods are the best, but I doubt that.

The reason why I think that trying out 2H methods for OH might be a good idea is that you can come up with really unique ideas. Let's use Roux again: If we have a close look, LSE is actually really good for OH. It's a low movecount way to finish the last remaining pieces without many algorithms, unlike methods such as ZZ which require hundreds to be that efficient. The only issue is that M moves seem to be undoable for OH, so anyone trying to create a good OH method would immediately discard that idea because that couldn't possibly be fast. With table abuse, however, LSE can reach 5-6 tps (source).
I'm not saying that the only good OH method can come from a 2H method, but rather that stuff you might not even consider can turn out to be revolutionary, so a lot that might seem bad should still be tried out.

In general good 2H methods have the lowest move count possible that's also easy to recognize, and guess what, that's also pretty good for OH. You obviously want to optimize the moves for OH, but F2L has always been good for OH, and LL variants have been developed and improved over the years.

For 2H, very high tps can be reached, so methods usually focus on ergonomics rather than movecount. To keep up the high tps, one also needs to have very fast recognition times. For OH, on the other hand, tps can't be as high and here, low movecount is more important than with 2H, while recognition times can be a bit longer due to not being required to keep up the high tps. That means that your average 2H method (like CFOP) won't be that great for OH.

IMO zz is way better than cfop for OH just because you do all your EO first and it really feels like a method developed for OH, and afaik it was developed for 2H.

While ZZ is probably better than CFOP for OH, it's still not ideal due to the constant z rotations. (Z rotations aren't that bad, but they still take some time and kinda interrupt your flow.)

 

[maticuber]

The reason why I think that trying out 2H methods for OH might be a good idea is that you can come up with really unique ideas. Let's use Roux again: If we have a close look, LSE is actually really good for OH. It's a low movecount way to finish the last remaining pieces without many algorithms, unlike methods such as ZZ which require hundreds to be that efficient. The only issue is that M moves seem to be undoable for OH, so anyone trying to create a good OH method would immediately discard that idea because that couldn't possibly be fast. With table abuse, however, LSE can reach 5-6 tps (source).
I'm not saying that the only good OH method can come from a 2H method, but rather that stuff you might not even consider can turn out to be revolutionary, so a lot that might seem bad should still be tried out.

I agree that LSE is really efficient, but ZZ can have really fast LL with just COLL+ 2gen EPLL, sure you can learn full ZBLL and it might be more beneficial in OH than it is in 2H.

For 2H, very high tps can be reached, so methods usually focus on ergonomics rather than movecount. To keep up the high tps, one also needs to have very fast recognition times. For OH, on the other hand, tps can't be as high and here, low movecount is more important than with 2H, while recognition times can be a bit longer due to not being required to keep up the high tps. That means that your average 2H method (like CFOP) won't be that great for OH.

I heard those same arguments more than 10 years ago when I started doing OH, and back in that era the 2H tps were lower than today's OH. It wasn't true back then, and it's not true today. People already have 6+tps with cfop in OH.

While ZZ is probably better than CFOP for OH, it's still not ideal due to the constant z rotations. (Z rotations aren't that bad, but they still take some time and kinda interrupt your flow.)

Yes, but it's way faster than doing any other move/rotation other than RU (or table abuse MU I guess).

 

[Deleted member 51471]

Waterman could be good, because after CLL, the cube is RrUM.

 

[brr_ya]

In my opinion, I don't think that the "optimal" OH has been invented yet. For example, TH solving had been around for much longer than OH solving. Even TH started with Corners First and Petrus because people just didn't have enough experience with the cube the invent the Cross or F2B. I mean, I know a lot of y'all are method developers, and would you guys have been able to think to make an cross from absolute scratch? Like how pre-historic humans made the wheel. It's simple to us but not to them. I argue that the same thing is happening now. All these so called "OH optimized methods" are just basically putting together things that are known to us like EO and CP. Like people didn't even know what EO was when it first started. I say that if people are as motivated as now to develop a new OH method, we'll discover something new like, whatever, and CFOP for OH will seem prehistoric similar to how we see people who hunt with bow and arrows.

 

[NevEr_QeyX]

In my opinion, I don't think that the "optimal" OH has been invented yet. For example, TH solving had been around for much longer than OH solving. Even TH started with Corners First and Petrus because people just didn't have enough experience with the cube the invent the Cross or F2B. I mean, I know a lot of y'all are method developers, and would you guys have been able to think to make an cross from absolute scratch? Like how pre-historic humans made the wheel. It's simple to us but not to them. I argue that the same thing is happening now. All these so called "OH optimized methods" are just basically putting together things that are known to us like EO and CP. Like people didn't even know what EO was when it first started. I say that if people are as motivated as now to develop a new OH method, we'll discover something new like, whatever, and CFOP for OH will seem prehistoric similar to how we see people who hunt with bow and arrows.

I'm not sure that that is possible because there already are methods that exist that have lower movecounts and are very difficult to understand. But they are not feasible because the amount of mental energy needed to use them is very high and nobody can think that fast.

 

[NevEr_QeyX]

I think ZZ-Roux is probably the best method for speedsolving because of the very low movecount using STM and the intuitive block building aspect.
with normal Roux coming in next, and CFOP being a close third with, ZZ dead last.

Edit: I meant to edit my original post oops.