Petrus Method

[shoot1510]

Oh no, not the "The world record is with Fridrich" thing again. If we keep thinking that, speedcubing will never get better. Eventually, people will have mastered Fridrich and there will be no getting better.

watch the first 20 seconds everything else is boring

I'm gettin a pain in the neck learning all F2L cases, Full OLL and PLL.

How does a kid or a person start cubing with Fridrich.
1. Kid buy a official Rubik's cube and read the instructions how to solve the Rubik cube. (LBL-Similar to Fridrich)
2. Then he/she want to get faster by going to google and search how to get faster at Rubik cube and learn more algs. (keyhole method, F2l, zz, oll, pll, finger tricks)
3. As time goes on, the kid will later on master the Fridrich method.

 

[Lord Voldemort]

Well, in Fridrich, all you do is place 4 edges.

It's not like you're not allowed to plan more than that. Many of the fast cubers look for the first ce-pair, too.

<RUL>

<R,U,L>

... the 800 you need for ZBLL with Fridrich

How did you calculate this? Anyway it's not ZBLL, I don't know if there's a name for it; usually it's called just 1-look LL.

Final note: EOLine actually needs less turns on average than a cross

How did you measure this?

ZZ F2L is optimal 26 turns ... For Fridrich, I think it's around 28 for F2L.

Would be a good idea to cite where these numbers are coming from. Solving the whole F2L step in one go optimally? Probably not, so how is it divided to sub-steps and how they are solved, etc.. I don't know how ZZ solvers do it.

The 4 ce-pairs in Fridrich can be solved in about 23 moves, see Stefan's Hume.

In order of quotes

1. I didn't account for that, I suppose. However, it could be done with ZZ as well.
2. I shall remember commas for the future.
3. I might be confused, but to get a 1LLL with pure Fridrich, the most well known way (or so I thought) was ZBLL. ZBLL uses 500 for 1LLL and then another 300 to orient edges while simultaneously forming and inserting the last F2L pair. Could someone correct/affirm me on this?
4. The place where I found it cited this for the cross.
this is the page for EOline.
I see now that I was comparing QTM for cross with HTM for EOline, my apologies for the misinformation. I'll edit the post.
5. I was also citing the same study. For what he calls "greedy optimality", F2L would take 23+6=29, as I was including Cross in the F2L. I have already showed you my source for EOline, and here's the post talking about ZZ blocks (I think you'll find the source trusworthy, considering it is from you )

Anything I missed?

 

[Cride5]

ZZ F2L is optimal 26 turns ... For Fridrich, I think it's around 28 for F2L.

Would be a good idea to cite where these numbers are coming from. Solving the whole F2L step in one go optimally? Probably not, so how is it divided to sub-steps and how they are solved, etc.. I don't know how ZZ solvers do it.

The 4 ce-pairs in Fridrich can be solved in about 23 moves, see Stefan's Hume.

I'm not sure if this is the 'proper' ZZ way to do it, but I solve the L & R 1x2x3 blocks in four stages. Two 1x2x2 blocks, and two 1x1x2 blocks. The order depends on the situation, but a 2x2 always comes before its connecting 1x2. I have no optimal figures for this strategy, but after an average of 10 solves I got 27.0 (not including EOLine). I'm sure the optimal figures would be much less, especially if each side was solved as one.

Johannes, where did you get the code for your online solver? I wouldn't mind compiling a few stats

 

[Lord Voldemort]

See my reply, it's the link before the

 

[Cride5]

Thx... so that gives 19.7176 moves for doing each side as one. I wonder why I'm using so many moves :confused: It would be interesting to see the optimal stats for the 4-stage strategy, but it might prove difficult to specify programatically since block selection depends on the situation.

 

[Lord Voldemort]

Optimality is HARD...
You have to blockbuild while setting up the next block.

 

[Johannes91]

3. I might be confused, but to get a 1LLL with pure Fridrich, the most well known way (or so I thought) was ZBLL. ZBLL uses 500 for 1LLL and then another 300 to orient edges while simultaneously forming and inserting the last F2L pair. Could someone correct/affirm me on this?

That method is called ZB. It's just like Fridrich, but instead of solving the last pair normally, LL edges are oriented at the same time (a.k.a. ZBF2L). Then the LL is the same as in Petrus and ZZ, and doing it in 1 look is called ZBLL. It requires 177 unique algs or 271 with inverses counted separately or 494 with mirrors counted separately, too.

So 493 + whatever ZBF2L needs is around 800. I thought you meant doing the Fridrich LL in one look without modifying F2L, which needs more than that (1212 or 2054 or something else).

5. I was also citing the same study. For what he calls "greedy optimality", F2L would take 23+6=29, as I was including Cross in the F2L.

Oh, ok. I blame the cubers who don't include it for confusing me.

The steps I used in the ZZ numbers are significantly longer than the ones in Fridrich F2L, so I'm not sure how meaningful comparing the numbers is.

Anything I missed?

Nope, thanks.

Johannes, where did you get the code for your online solver? I wouldn't mind compiling a few stats

I wrote it, but most ideas are from Jaap's and Kociemba's sites. If the stats you're thinking of are similar to the solvers on that page (not something like solve a 4x4x4 maze cube optimally), just let me know and I'll generate them.

 

[Nukoca]

Are there algs written out for something like EJF2L, but for COLL? So you can have a corner unoriented in the F2L and get it fixed during the COLL?

 

[Cride5]

Johannes, where did you get the code for your online solver? I wouldn't mind compiling a few stats

I wrote it, but most ideas are from Jaap's and Kociemba's sites. If the stats you're thinking of are similar to the solvers on that page (not something like solve a 4x4x4 maze cube optimally), just let me know and I'll generate them.

... OK this is a big ask, so no worries if you don't have time! On the ZZ 1x2x3 page would it be possible to include options for breaking it down into 1x2x2 + 1x1x2 blocks. As a ZZ cuber it would be very nice to be able to compare my blockbuilding with the optimal solutions.

Optimal blockbuilding in four stages is certainly a realistic prospect for speedsolvers so the stats for this would hopefully provide a realistic comparison to Fridrich F2L.

... apologies Petrus users, we seem to have veered wildly off-topic

 

[Vryon]

Alright, "Petrus" and "Roux" is the same. You still can solve the 3x3x3 rubiks cube with both ways. >__>

 

[Lord Voldemort]

The steps I used in the ZZ numbers are significantly longer than the ones in Fridrich F2L, so I'm not sure how meaningful comparing the numbers is.

Could you elaborate?
I don't understand what you mean by steps.
Are you talking about the furthest number of turns it looks ahead to find a solution?
One more question: is Hume finished/released yet?
It sounds like a useful piece of software.

 

[brunson]

5. I was also citing the same study. For what he calls "greedy optimality", F2L would take 23+6=29, as I was including Cross in the F2L.

Don't use the abbreviation F2L for that. F2L has a very specific meaning in the context of CFOP. Use either "Cross+F2L" if you mean cross plus Fridrich F2L or "first two layers" for anything else.

 

[Nukoca]

Can we move this discussion to a different thread? When people click on this thread, they expect to read a discussion about Petrus, not ZZ or ZB!

 

[Cubemaster37]

I understood the Petrus Method but which Algorithms you are using to finish the last layer so fast?

 

[soccerking813]

Most people use corner OLL and then PLL I think. Or COLL and EPLL.

 

[JLarsen]

I use oll pll. Those oll's are so fast, and pll is decent as well.

 

[James Kobel]

You are all wrong, and frankly you are all being idiots. The term "best" is completely opinion, and no one's opinion is right. If you were going for something like which method is fastest, that would be fine, but this has no fact involved. You can't shange someone's opinion without fact, and the only facts you are bringing up are really just opinions, such as "Petrus last layer is harder" or something like that, it all depends on who you are and what you use. So shut up.

 

[shoot1510]

You are all wrong, and frankly you are all being idiots. The term "best" is completely opinion, and no one's opinion is right. If you were going for something like which method is fastest, that would be fine, but this has no fact involved. You can't shange someone's opinion without fact, and the only facts you are bringing up are really just opinions, such as "Petrus last layer is harder" or something like that, it all depends on who you are and what you use. So shut up.

For example: If Gives you hell is the top #1 song at MTV, it their opinion only.

 

[Lord Voldemort]

5. I was also citing the same study. For what he calls "greedy optimality", F2L would take 23+6=29, as I was including Cross in the F2L.

Don't use the abbreviation F2L for that. F2L has a very specific meaning in the context of CFOP. Use either "Cross+F2L" if you mean cross plus Fridrich F2L or "first two layers" for anything else.

Can't it also mean first two layers though?
I realize that it's a step in Fridrich, but most people I've seen use it describe first two layers.

 

[MistArts]

When doing FM solves, in the back of your mind you're always thinking about move efficiency - ie: how much closer does each move take you to your goal. I believe the moves made during EOLine do a lot more to help the rest of the solve than the equivalent number of moves used during the Fridrich cross. EO helps you during F2L, with move ergonomics and lookahead. It also helps you on LL by pre-orientation of the LL edges, leading to easier/more efficient LL options.

It seems that ZZ should be better than Fridrich in theory, but we've yet to see that proved in practice. All the more reason to learn it, and learn it good

In FM, you try to orient pieces independently. Your moves should solve a section of a cube. Orientation doesn't solve a piece, unless it's already permutated, and you usually don't "pure" orient pieces. What you should do in FM is to permute in the correct orientation from any starting orientation. So with ZZ, it doesn't help your moves, it restricts them. I'm not saying Fridrich is better at FM. It has restrictions too. Freestyle block-building gives less restrictions to these 'methods'. Of course, FM is different than speed.