Semi F2L Roux Variant

[Dafe]

This is my first post, so sorry if this is in the wrong section. Didn't seem like a beginner topic, and I couldn't find any threads about it here.

For context, I went straight from an edges first beginner's method to Roux, and currently average in the high 20s (my turn speed isn't great, and I don't know CMLL). I don't really have any experience with other speedcubing methods, but I'm familiar with some conceptually.

Anyway, I was wondering if any of the Roux users here, who also know Fridrich, have any insight on this.

One thing that's always bothered me with Roux is that steps 1 and 2 are separate, even though it seems like it's more flexible to do them simultaneously.

i.e. I solve with white on my left, yellow on my right, and red on the bottom, so I could set up my white/red edge and yellow/red edge opposite each other on the bottom, like just 2 out of 5 parts of solving a red cross, and then do F2L to work on both sides at the same time with whatever's convenient.

I've tried this a little, but I don't have an F2L background, so it slows me down a little.

Have any Roux/Fridrich users tried this much? Does it help solve times?

 

[ostracod]

Welll, I'm a ZZ solver, and for me it is considerably faster to solve both blocks at the same time. In the ZZ method, there is a separating line between the blocks, unlike Roux, and all the edges are oriented (so you make only L, U, and R moves). The logistics of block building in ZZ and Roux are probably different, but I would venture to say that it would be more efficient in Roux as well to build blocks interchangeably. The reason is that you can construct parts as you SEE them, instead of searching for parts, which can take up precious time. The only downside I see is that building parts of blocks on both sides restricts your freedom a bit, but it seems that this would only be a problem for low move count competitions.

...If a Roux solver has some argue against/for this, please make it!

 

[Lofty]

There is a big difference between Roux and ZZ and that is how inspection is used. (Well on top of the obvious differences haha) In Roux you are supposed to see the entire first block in inspection. So because of that its better to just do the one block at a time.

 

[soccerking813]

Finding all six of those pieces and knowing how to place them in 15 seconds is very hard to do I think.
And if this makes it easier for you to go about making the blocks, then use it. It makes it quite a bit easier for me as well.

 

[Kirjava]

Personally, less block building freedom slows me down.

 

[BigGreen]

Finding all six of those pieces and knowing how to place them in 15 seconds is very hard to do I think.

its really not that hard

 

[Dafe]

It's probably easier than doing an extended cross.

In any case, how does time for step 1+2 for Roux generally compare to cross+F2L, for equivalent total solve times?

 

[Johannes91]

It's probably easier than doing an extended cross.

It should be. XCross is 6 pieces, 1x2x3 block 5. Optimal average for a fixed 1x2x3 is just 6.70 moves, but for XCross it's 7.98.

 

[deadalnix]

Step 2 in roux method can be done using rRU and it's why it's a differents step.

Obvously, you should do step 1 and 2 simultanously in FM.

 

[somerandomkidmike]

I'm actually gonna try this. I use waterman.

 

[Jhong253]

Well, roux can be quite difficult at first if you are totally unfamiliar with block building... I'm so used to the fridrich way that I see the fridrich f2l pair in less than a second whereas I still have trouble getting that first 1x2x3 block for Roux unless I have a lot of time...

 

[Dafe]

Does anyone know if there are any statistics available for Roux step 1+2 and Fridrich Cross+F2L times for like equivalent solve times?

 

[Johannes91]

Does anyone know if there are any statistics available for Roux step 1+2 and Fridrich Cross+F2L times for like equivalent solve times?

What exactly do you want? I can't understand that sentence.

Anyway, I don't think solve times for different methods can be meaningfully compared. They depend so much more on the cuber than the method, and even the fastest users of each method are not "perfect" yet.

For move counts there are some statistics, but it's of course just one of the many factors in speedsolving.

 

[Lofty]

Does anyone know if there are any statistics available for Roux step 1+2 and Fridrich Cross+F2L times for like equivalent solve times?

What exactly do you want? I can't understand that sentence.

Anyway, I don't think solve times for different methods can be meaningfully compared. They depend so much more on the cuber than the method, and even the fastest users of each method are not "perfect" yet.

For move counts there are some statistics, but it's of course just one of the many factors in speedsolving.

I think what he means is, say you want a 15s average you should have F2L around 10seconds what would the equivalent step 1+2 time be? Tho I agree thats not much of a comparison to make, something like move count is better.

 

[Dafe]

Yeah. Lofty's interpretation is what I meant.

The impression that I get is that Roux is more efficient in terms of move count, but average solve times are by and large slower, so in the context of speed solving, that doesn't matter that much. Algorithm ergonomics and case recognition speed and lookaheadability are all factors that have effects on execution time that aren't explicitly tied to move count.

I know Roux hasn't been worked on as much as Fridrich, which is why I'm not asking which method is faster, or which has better potential. I'm only interested in the proportions.

Since Roux step 1+2 can be executed as like half a cross + F2L, if F2L isn't slower, there's a problem with Roux steps 1 and 2 being separate steps.

The issue I feel is that while you have more freedom with what kind of moves you can do to build your blocks, you have fewer options about what cubies you can work with, which hurts recognition and lookahead since the ones you can use are less likely to be in view at any given time.